Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2007-04-18

00:01:43 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
00:03:18 Quit davina (Remote closed the connection)
00:03:37markunroolku: a bricked Gigabeat could be fixed with a jtag interface
00:03:43Rifti don't find the link to login into the wiki
00:04:13sslashesjust click an edit link or whatnot and it will ask for credencials
00:04:33roolkumarkun: from his description the (rom-) bootloader still seems to work
00:04:51 Quit amiconn (Read error: 110 (Connection timed out))
00:04:51 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
00:05:49 Nick jhulst_ is now known as jhulst (
00:06:13roolkuI just don't understand what he has done and why he would want to run fw.exe
00:07:44gigapeteanyone know where i can get the system files for a gigabeat f40?
00:08:50markungigapete: in the wiki you will find a zip with a fake gbsystem folder
00:09:02markunif you just use rockbox you can use that one
00:09:11markunit will speed up booting too
00:09:58 Join daniel2023 [0] (i=d0b4e09e@gateway/web/cgi-irc/
00:11:51dionoeayay! it's rendering text from the srt file
00:11:55 Join lids [0] (i=lds@gateway/tor/x-3858bab56c244c1b)
00:16:27 Join Hammer89 [0] (
00:18:23 Quit lids (Remote closed the connection)
00:18:50 Quit obo ("bye")
00:20:09 Join lini [0] (i=pugsley@
00:21:39markundionoea: watching anime with subs? :)
00:22:29dionoeano ... finished writing a logo erase video filter for VLC so i though i'd have fun on rockbox a bit :)
00:22:56 Quit idnar (Connection timed out)
00:23:32markunhow weird, I just received SPAM with a subject in esperanto!
00:24:12markunthe rest of the email is computer generated english, and a picture is attached with the actual spam
00:25:45 Join lids [0] (i=lds@gateway/tor/x-787fe189b9d06ce5)
00:27:26 Quit bluebrother ("leaving")
00:31:15 Join brenda [0] (
00:33:24 Join haTem [0] (
00:33:57 Nick _pill is now known as pill (
00:34:28dionoealinuxstb: is curr_time the time i should use to position the subs ? or eta_video ? (or something else?)
00:34:31 Quit brenda (Remote closed the connection)
00:37:37dionoeahum, curr_pts would be better in fact
00:37:44linuxstbdionoea: I don't know any more - best to ask jhMikeS
00:38:01dionoeawhat unit is that using btw ?
00:38:29 Join webguest02 [0] (i=c874c1b8@gateway/web/cgi-irc/
00:38:32 Quit JdGordon ("Konversation terminated!")
00:38:47linuxstbI think the PTS values are divided by 2 - so they are 45KHz. But the main clock is the DAC - 44.1KHz.
00:39:08webguest02I have a big trouble!!
00:39:09linuxstbSo I think all times are converted into audio samples.
00:39:19 Quit funky ("leaving")
00:39:41dionoeaso 44.1 kHz ?
00:40:58webguest02i change battery type of my ipod nano and every i try to turn on the ipod its appear very low battery!
00:41:05webguest02please help me!
00:42:01linuxstbwebguest02: Charge it?
00:42:21 Quit entheh ("^~")
00:42:52webguest02i try but wen i disconnect it appear again @very low battery@
00:43:23linuxstbHave you reset it (hold MENU+SELECT) ?
00:43:59webguest02i try restoring with itunes but dont work
00:44:05dionoealinuxstb: any idea if the time starts at 0 ? or is it some internal clock used by the stream with a random offset ?
00:44:23linuxstbdionoea: It will most likely start at the first PTS of the first audio frame.
00:44:40linuxstbWhich can be anything...
00:44:40dionoeais that value stored somewhere ?
00:45:09linuxstbI used to store it, but I think jhMikeS deleted it. Search for "first" in mpegplayer.c
00:45:42dionoealooks like it's gone :(
00:45:42 Quit daniel2023 ("CGI:IRC (EOF)")
00:45:43 Quit mirak (Remote closed the connection)
00:45:46linuxstbwebguest02: What do you mean by "change battery type" ?
00:45:48 Quit moos ("Glory to Rockbox")
00:45:52dionoeaI'll store one my self (on the first call to the function)
00:46:42 Quit billytwowilly (Remote closed the connection)
00:47:26 Join billytwowilly [0] (
00:47:41 Quit lee-qid ("aufwiederbyebientotsayonara")
00:48:28webguest02in rockbox
00:48:35haTemhey, my iriver h320 only powers on when plugged into a power supply, and I believe this started after I upgraded rockbox to one of the recent daily builds (I had been using a build from sometime in october)... has anyone experienced something like this?
00:48:41dionoeaoops, Prefetch abort. That sounds bad
00:48:43 Join stripwax [0] (
00:49:06haTemI've tried multiple firmware versions, and I've tried installing the oldest daily build I could find on the website (3/16)
00:49:29webguest77haTem: there's a config option for usb charging
00:49:32webguest02i accidentaly turn the battery type in to another
00:49:45stripwaxDoes the new customizable icons code mean that icons on 16-bit targets are always displayed in black regardless of the foreground color (=> custom wps with black backgrounds are need required to now have custom icons) ?
00:50:47webguest02no one can help me_
00:51:21 Quit inversions (Read error: 113 (No route to host))
00:52:19haTemwebguest77: "Charge During USB Connection" is set to "No", could that be causing it?
00:52:56webguest02i dont know
00:52:57webguest77that will stop it charging via usb... is that what you meant to start with?
00:53:04webguest02i cant restore my ipod now
00:53:37webguest02i cant start with rockbox
00:53:49haTemwebguest77: no, what I mean is that if it's not connected to a power supply, it refuses to turn on (even though it appears to be fully charged)
00:54:58webguest77haTem: ah, sorry I misunderstood - I don't know why that should be the case
00:55:02webguest02it turn on but i have to connect and reboot about 6 times
00:55:41linuxstbwebguest02: Do you mean the Battery Capacity setting?
00:56:11webguest02yes battery capacity setting
00:56:15haTemwebguest77: heh, I'd think it was just some freak coincidence and that the battery is just dead (unrelated to the rockbox upgrade), but my friend is having the same issue with his recently purchased h320
00:56:36linuxstbwebguest02: That setting does nothing at all to your ipod - it's just used to help Rockbox estimate the remaining runtime to display.
00:57:15webguest77haTem: sure - for what it's worth my h320's fine...
00:57:24linuxstbwebguest02: It simply sounds like your battery is low - charge it, wait, and reboot with MENU+SELECT.
00:57:59webguest02i charge it for 6 hours and it dont charge
00:58:20ashesif i email my sister an ogg vorbis file, will she have any trouble playing it in windows?
00:58:28stripwaxSo (I would have thought) the default icons should be 1bpp on 16bit targets, so that they DO show up in foreground color. but that doesn't seem to be the casse
00:58:32 Join idnar [0] (n=mithrand@unaffiliated/idnar)
00:58:51stripwaxashes - is that a Rockbox question?
00:59:14haTemwebguest77: are you using a recent build of rockbox?
00:59:22haTemI suppose I should try fiddling with the battery
00:59:35linuxstbstripwax: I agree...
00:59:39webguest77haTem: today's or maybe yesterday's now
00:59:40haTemalso, what firmware are you using :P
00:59:53ashesif i email my sister an ogg vorbis, from my iriver which runs rockbox, will she have any trouble playing it in windows?
01:00:04webguest02last firmware
01:00:22linuxstbashes: She can copy it to her DAP running Rockbox and play it...
01:00:23dionoealinuxstb: looks like get_playback_time() was what i was looking for (it changes at about 44kHz and starts at 0)
01:01:08webguest02so sad...
01:01:19webguest02but tks a lot
01:01:52stripwaxashes: sounds like a Windows question. maybe, the answer will be on
01:02:06webguest02i go to put my ipod into a micowave
01:02:25 Join smolyn [0] ( says she'll need additional software
01:02:41stripwaxashes - well it depends what software she already has ....
01:02:53stripwaxwebguest02 - what is the issue?
01:03:18linuxstbstripwax: ipod nano refusing to charge
01:03:27dionoeawebguest02: make sure that you make a movie of the ipod in the microwave
01:04:17webguest02and ipod cant run nomore linux and rockbox
01:04:55stripwaxwebguest have you tried leaving it to fully charge via a mains adapter (not usb)?
01:05:18webguest02yes but the problem im the same
01:05:57 Part Rift
01:06:31webguest02please help me
01:09:00stripwaxwebguest02 - so it doesn't even turn on? no icons on the screen saying charging? you've tried resetting it?
01:11:05webguest02yes it turn on
01:11:58webguest02but in the apple os dont show the battery and wen i disconnet it it turn of
01:12:23***Saving seen data "./dancer.seen"
01:12:30webguest02wath i do_
01:13:04webguest02i even restoe my ipod with itunes i i have the same problem
01:14:53dionoeaCool. I've got srt subs working :)
01:14:58webguest02any of you have any idea?
01:15:05dionoeaI'll clean it up, add some error checking and post a patch here tomorrow
01:17:01linuxstbdionoea: What's the cost? i.e. how much slower is it?
01:17:50dionoeai've lost 1 or 2 fps here (but i'm rendering debug output at the time)
01:17:56dionoeaSo i'll see once it's cleaned up
01:18:02linuxstbOn which target?
01:18:06dionoeaipod video
01:18:13dionoea(worst of all :) )
01:18:29linuxstbSo down from 10fps to 8 or 9 (with sync enabled) ?
01:18:30dionoeaand i'm rendering debug output at 4HZ instead of 1HZ like in the normal plugin
01:18:40dionoealinuxstb: i guess so... let me try
01:19:43webguest02please help me
01:19:50dionoea8.5 to 9.5 fps during the first 20 secs of elephant dreams
01:20:01stripwaxwebguest02 - sorry, I can't help, I'm not sure what you mean when you say "in the apple os dont show the battery". It turns on but the Apple OS doesn't work? It doesn't sound like a rockbox issue, perhaps your ipod is faulty
01:20:02dionoeaso performance stays about the same
01:20:14dionoeaI've really got to get some sleep now. I'll post a patch tomorrow
01:20:16dionoeagood night
01:20:17linuxstbWell, the first 20 secs is the easiest to decode...
01:20:28 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
01:20:32dionoeaI'm at 8.9 now
01:20:41dionoeaseems to be pretty stable arround 9 fps
01:21:08dionoea(my timing looks a bit screwed ... based on real time and not stream time it seems)
01:21:15webguest02wen i go to te apple os the batery icon dissapears
01:21:25pixelmawebguest02: there were a lot of reports that ipods, especially nanos, behave weird on low battery. Most of what I read says: try charging it for a longer time, do the reset and it'll get back to normal. See as a good example (given that there is no other hardware issue of course)
01:22:29dionoeaah no, listening to the audio at the same time shows that it's synced
01:22:33dionoeaawesome :)
01:22:36 Quit ompaul ("there is no fancy quit message - just this")
01:22:39*dionoea 's off to bed
01:25:40 Quit PaulJam (".")
01:28:49webguest02i have the problem of pianofreq
01:29:10webguest02and seems like he dont fix hes ipod
01:30:20 Quit roolku ()
01:31:31 Quit Thundercloud (Remote closed the connection)
01:33:31 Join Thundercloud [0] (
01:39:58 Quit haTem ("i quit.")
01:45:26 Join barrywardell [0] (n=barrywar@
01:54:02 Join elinenbe__ [0] (
01:56:56 Quit elinenbe (Read error: 110 (Connection timed out))
01:56:57 Nick elinenbe__ is now known as elinenbe (
01:57:46 Join pearldiver [0] (
01:59:46 Join Yono [0] (
02:05:50pearldiveris anybody using the new dict patch?
02:05:52 Quit webguest02 ("CGI:IRC (Ping timeout)")
02:09:11 Quit barrywardell ()
02:10:44 Join ceaser_ [0] (
02:11:23 Part Domonoky_
02:13:31 Quit perldiver (Read error: 110 (Connection timed out))
02:39:24 Quit midgey ()
02:50:29 Join Nico_P [0] (
02:54:54 Quit jhMikeS (Read error: 60 (Operation timed out))
02:55:44 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
02:59:59 Join webguest87 [0] (i=cf60f3c3@gateway/web/cgi-irc/
03:00:04webguest87hi everybody
03:00:26webguest87i have got one question - is rockbox gonna work on my ipod (80GB 5.5G) ??
03:03:32Hammer89unless it's listed here I don't think so...
03:04:56 Quit webguest87 (Client Quit)
03:05:19 Part pixelma
03:05:54 Join webguest80 [0] (i=89a5f07d@gateway/web/cgi-irc/
03:12:26***Saving seen data "./dancer.seen"
03:12:30 Quit webguest80 ("CGI:IRC (EOF)")
03:19:23 Join N4z3r_0f_sun [0] (
03:19:46N4z3r_0f_sunhello? is this the rock box channel?
03:20:33midkayyes, it is.
03:21:06N4z3r_0f_sunok can you help me i want to use my ipod mini with it ok?
03:21:27N4z3r_0f_sunbut im totally lost and i have no clue where to start
03:21:39midkayfollow the manual instructions. if you have a specific problem or question then you can ask.
03:21:46N4z3r_0f_sunkk ty
03:21:46midkayyou don't need permission to ask questions in this channel.
03:22:05 Part N4z3r_0f_sun
03:27:43 Quit spiorf ("Read error: 110 (Connection timed out)")
03:31:09 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
03:31:14 Join miepchen^schlaf [0] (
03:34:33 Quit Nico_P (Remote closed the connection)
03:40:18 Quit Thundercloud (Remote closed the connection)
03:41:21 Join TrueJournals [0] (
03:42:05 Quit webguest77 ("CGI:IRC (EOF)")
03:43:20 Join GRaTT [0] (
03:45:09GRaTTHi need some help reading files for a playlist converter on sansa
03:46:35GRaTTalready have rockbox playlist to sansa Original firware palylist
03:46:54GRaTTtrying to do a sansa OF playlist to rockbox m3u
03:47:16GRaTThaving trouble due to null bytes in .pla playlist files
03:48:57 Quit Guile (Read error: 104 (Connection reset by peer))
03:50:59 Quit lids (Remote closed the connection)
03:52:26 Quit miepchen^schlaf (
03:52:26 Quit smolyn (
03:52:26 Quit billytwowilly (
03:52:26 Quit jhulst (
03:52:26 Quit desowin (
03:52:26 Quit alberink (
03:52:26 Quit GodEater (
03:52:26 Quit jaebird (
03:52:26 Quit hcs (
03:52:26 Quit sando (
03:52:26 Quit gtkspert (
03:52:26 Quit GodEater__ (
03:52:26 Quit z0de (
03:52:26 Quit secleinteer (
03:52:26 Quit dan_a (
03:52:35NJoinmiepchen^schlaf [0] (
03:52:38 Join lids [0] (i=lds@gateway/tor/x-2230d8740e9e2f6b)
03:53:09NJoinsando [0] (
03:53:19 Join johnnyoc3 [0] (
03:53:47 Join Llorean [0] (n=Llorean@rockbox/administrator/Llorean)
03:55:23NJoingtkspert [0] (
03:57:02 Join borisyeltsin [0] (
03:57:21 Join alberink_ [0] (
03:59:13johnnyoc3anyone know some good deals on sansa players? im looking to get one
04:00:08NJoinsecleinteer [0] (
04:01:34NJoinsmolyn [0] (
04:01:34NJoinbillytwowilly [0] (
04:01:34NJoinjhulst [0] (
04:01:34 Join desowin [0] (n=desowin@unaffiliated/desowin)
04:01:34NJoinalberink [0] (n=alberink@
04:01:34NJoinGodEater [0] (
04:01:34NJoinjaebird [0] (n=jae@
04:01:34NJoinhcs [0] (n=agashlin@rockbox/contributor/hcs)
04:01:34NJoinGodEater__ [0] (
04:01:34NJoinz0de [0] (
04:01:34NJoindan_a [0] (n=dan_a@
04:01:35***Server message 505: 'logbot- :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
04:02:14 Quit z0de (Connection reset by peer)
04:03:04GRaTTanybody good with text file reading in rockbox
04:03:15 Quit desowin (Excess Flood)
04:04:36 Join jhulst_ [0] (
04:04:50TrueJournalsGRaTT: What do you mean?
04:04:54johnnyoc3what do you mean "good with"
04:05:17 Quit alberink (Read error: 110 (Connection timed out))
04:06:20 Quit jhulst (No route to host)
04:06:38 Nick jhulst_ is now known as jhulst (
04:09:56GRaTTTruejournals: I am working on a playlist converter for sansa
04:10:30GRaTTTrueJournals: I need to read in the OF playlist file and it has null bytes between each char
04:11:07TrueJournalsah, so you're saying coding a plugin for rockbox?
04:11:37GRaTTTrueJournals: yes, I already have rockbox playlist to OF playlist
04:12:08TrueJournalsAh, OK, not me then :-p
04:12:11GRaTTjohnnyoc3: anybody that can helpo me read these sansa OF playlist files (.pla)
04:12:37TrueJournalsGood luck though
04:12:37johnnyoc3sorry i cant help with that :(
04:14:11GRaTTanybody here work with sansa playlists at all
04:16:16LloreanThere's usually not much care for the original firmware.
04:19:01 Quit gtkspert (Remote closed the connection)
04:19:03 Quit mbr (Read error: 60 (Operation timed out))
04:19:29 Join gtkspert [0] (
04:19:38rotatornull bytes between each character? it sounds like it's some unicode encoding with >8 bits per symbol
04:22:24 Join webguest22 [0] (i=d360b527@gateway/web/cgi-irc/
04:24:30 Part TrueJournals
04:24:52 Quit Soap2 (Read error: 110 (Connection timed out))
04:25:07 Quit billytwowilly (Connection timed out)
04:27:15johnnyoc3anyone had issues with the cap WPS freezing up on them on a 30gb ipod?
04:27:49 Quit webguest22 ("CGI:IRC (Ping timeout)")
04:28:06GRaTTLlorean: I know , I started this before sound, alway some reason to use OF though.
04:29:35 Quit ceaser_ (Read error: 104 (Connection reset by peer))
04:34:43GRaTTrotator: yes something like that. Just having trouble reading the file due to the null bytes
04:35:15pearldiverLlorean ha, too bad it wasnt a champaign after all!
04:35:15GRaTTrotator: using the text editor in RB just the first char displays
04:35:31GRaTTrotator: using the text viewer in RB displays the whole file
04:35:52 Join saratoga [0] (i=98039b62@gateway/web/cgi-irc/
04:36:00Lloreanpearldiver: I got it in better light, it is.
04:36:07LloreanIt's apparently just a very subtle color
04:36:16pearldiverit is
04:36:17 Join ozgur [0] (i=a46b7415@gateway/web/cgi-irc/
04:36:52pearldiveralso the outer rim is completely different to the touch to F20 and F60
04:37:13LloreanOn F40s?
04:37:14 Quit saratoga (Client Quit)
04:37:39 Join saratoga4 [0] (i=98039b62@gateway/web/cgi-irc/
04:37:49pearldiverLlorean yes
04:38:03LloreanWell, I've never touched an F20 or F60. :)
04:38:25pearldivermy black F20 is extremely polished
04:38:28saratoga4GRaTT: checked unicode.c?
04:38:36 Join webguest89 [0] (i=89a5f07d@gateway/web/cgi-irc/
04:38:45saratoga4i'd guess theres a function in there
04:39:20pearldiverF40 almost feels like it has a texture to it
04:39:20 Quit ozgur (Client Quit)
04:40:10 Join Topy44 [0] (
04:42:16 Join aliask [0] (
04:43:44saratoga4wow amazon added 20 dollars to the 8GB Sansa price
04:43:54saratoga4but I think the actually lowered the price on the 6GB
04:44:03saratoga4Sandisk pricing makes no sense at all to me
04:44:24 Quit johnnyoc3 ("ChatZilla [Firefox]")
04:45:30GRaTTsaratoga4: I will look now thanks
04:46:58 Join JdGordon [0] (i=82c20d66@gateway/web/cgi-irc/
04:47:56saratoga4JdGordon: did you ask about the Sansa emulator the other day on IRC?
04:49:19JdGordonI did
04:49:32JdGordonI was actually planning on chatting with you this arvo about it :)
04:49:53JdGordondo you know if toni has got anywhere with the cop yet?
04:50:26saratoga4last i heard he'd given up in order to work on the sansa itself
04:50:43saratoga4i don't know what the final status of the emulator was, i haven't had a chance to try his latest version
04:51:01saratoga4plus you need a complete sansa disk image to do so, and I don't have such an image handy
04:51:39 Quit JdGordon (Client Quit)
04:51:43saratoga4the version i tried could navigate on pre-cop builds of rockbox and fill the PCM buffer if you tried decoding mp3s
04:52:53GRaTTsaratoga4: I am looking at the ut8decode function in unicode.c. can I even call these from a plugin?
04:55:01saratoga4should be able to
04:55:09rotatorGRaTT: all the functions in the struct plugin_api in \apps\plugin.h can be used in plugins
04:55:37rotatorutf8decode is included among them
04:56:26 Join JdGordon [0] (i=82c20d66@gateway/web/cgi-irc/
04:56:34JdGordonsaratoga4: sorry, comp problems
04:56:41saratoga4did you miss my reply?
04:56:50saratoga4"the version i tried could navigate on pre-cop builds of rockbox and fill the PCM buffer if you tried decoding mp3s"
04:57:15JdGordonhmm... im more interesetd in the OF..
04:57:16saratoga4i thought it would be useful for profiling code on ARM
04:57:26saratoga4since you can intercept all the memory accesses and log them
04:57:28 Quit Topy (Read error: 110 (Connection timed out))
04:57:41 Quit JdGordon (Client Quit)
04:57:42saratoga4it worked on the bootloader, but crashed in teh OF due to lack of good COP support
04:59:15 Join JdGordon [0] (i=82c20d66@gateway/web/cgi-irc/
04:59:22JdGordon3rd time lucky :p
04:59:50GRaTTsaratoga4: rotator : thanks. It looks like viewer.c uses the utf8decode function to read the pla files. I will try.
05:00:49JdGordonsaratoga4: do you know how to compile the emu? i want to fiddle with it and try to see what the OF does differently if a sd card is presenty
05:01:05GRaTTJdGordon: Hello, you left a reply on my m3u2pla converter at flyspry
05:01:12saratoga4you need VS6
05:01:13JdGordongratt: hey, yeah, the pla is unicode which is why the nulls are there..
05:01:30GRaTTJdGordon: I made some of the changes you suggested
05:01:33saratoga4I tried in vain to get the free VS package to compile it, but even with teh SDK installed, it claimed it couldn't find stuff
05:01:52saratoga4though theres almost no Windows specific code at all. so if you have any c++ background, you could port it pretty easily
05:01:55GRaTTJdGordon: unsure of mutlti-screen api for display
05:01:58JdGordonGRaTT: I saw the comment, havnt looked at the code yet though
05:02:28GRaTTJdGordon: trying now for pla->m3u
05:02:31JdGordonI can make it looks nice for you if you need (I dont tihnk it needs a display anyway...)
05:03:02saratoga4the annoying thing was figuring out what files it actually needed, and the required files had a habbit of changing between releases, so I usually used the "find" tool to look for file paths in the source code :)
05:03:08JdGordonsaratoga4: I installed vs6 yesterday so ill see how I go... does it need the updated windows sdk?
05:03:24saratoga4I don't think so
05:03:44saratoga4I had the SDK installed, but i don't think it needed it
05:03:55JdGordonok, great
05:03:58saratoga4though i know nothing about win32
05:03:59JdGordonmakes things easier
05:04:06 Quit Hammer89 (Read error: 54 (Connection reset by peer))
05:04:15GRaTTJdGordon: have not coded C in 7 years, I am very rusty. Is there better (faster) way for display
05:04:19JdGordonGRaTT: what was the FS #?
05:05:06saratoga4but yeah, if you want to work on it, that'd be great, I really think we could use a good arm emulator since it seems nearly all the rockbox targets from here on will be arm
05:05:17GRaTTJdGordon: JdGordon FS# 6884
05:05:51pearldiverwhy the text editor has empty menu items?
05:06:14 Quit linuxstb (Read error: 113 (No route to host))
05:06:20Llorean pearldiver: Spacing (I'm not fond of it)
05:06:36pearldiverany reason for that?
05:06:38 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
05:06:43GRaTTJdGordon: if display is on, each line is displayed it takes 15 min for 500 line playlist if off 1 second
05:07:59GRaTTJdGordon: display was really for debug during developement
05:08:03JdGordonGRaTT: you can do a splash with 0 timeout so it doesnt halt the proccess....
05:08:45JdGordonwhat you want is something like FOR_NB_SCREENS(i) { rb->screens[i]->lcd_puts(0,0,"some text"); }
05:09:18JdGordonFOR_NB_SCREENS(i) { rb->screens[i]->lcd_puts(0,0,"some text"); rb->screens[0]->update(); }
05:09:50GRaTTJdGordon: doing this multiple times would be faster?
05:09:59JdGordondont we have a ascii->utf8 function?
05:10:19GRaTTJdGordon: there is an utf8decode
05:10:23JdGordonwell, the way you have the splsh is it stops the proccess for 1 sec
05:11:07GRaTTJdGordon: I will try with 0 time out first for simplicity
05:11:54GRaTTJdGordon: I am looking at implementing the utf8decode as in viewer.c to read in the pla file
05:12:30***Saving seen data "./dancer.seen"
05:12:54JdGordonif its there already you can add it to the plugin api (plugin.[ch]) so you can use it without recoding it
05:13:30 Quit saratoga4 ("CGI:IRC")
05:15:46GRaTTJdGordon: can I call it directly?
05:16:33JdGordonno, ou have to add it to the api
05:16:40JdGordonthen you call it like rb->utf8decode()
05:17:55GRaTTJdGordon: it is called like that in viewer.c. Must be in the plugin api already?
05:18:44JdGordonoh, then yes
05:20:47GRaTTJdGordon: I just looked and do not see any info in PLUGIN_API
05:21:14JdGordonif viewer.c is calling it like rb->utf8decode then your fine to do the same
05:21:44 Quit borisyeltsin (Remote closed the connection)
05:22:23 Join billytwowilly [0] (
05:22:27GRaTTJdGordon: OK, thanks Thats more than enough to keep me busy for a while
05:32:39 Join toffe82 [0] (
05:34:18 Quit GRaTT ("using sirc version 2.211+KSIRC/1.3.12")
05:38:59 Join hachi [0] (
05:41:13hachiso... uh... the 5.5g thread has been locked for 2 months now... does anyone even remember about that?
05:42:03LloreanI do.
05:42:05LloreanAnd it's STILL TRUE
05:42:36LloreanDo you think we're going to finish it, and then forget to tell you?
05:42:55hachino, I think it's just fallen to the wayside and will never get done
05:43:05LloreanAnd what if it has?
05:43:26LloreanNobody else is volunteering to rewrite the Rockbox FAT32 code.
05:43:58LloreanThere is basically one person interested in working on it. And it's a job that doesn't split well. Either he'll get it done, if/when he has time, or he won't.
05:44:09LloreanUnless someone else comes along and volunteers, that's all you've got.
05:45:35LloreanSo, facing that knowledge, what did you want?
05:45:56*joshin suggests that hachi find out who the dev is and bribes them
05:45:57 Quit JdGordon ("CGI:IRC (EOF)")
05:46:13hachinothing in particular... do you want people to just shut up and never say that they still would love to have support for the device?
05:46:18Lloreanjoshin: Money doesn't make the real world and family go away.
05:46:44Lloreanhachi: I want them to stop opening new threads asking for status updates. I want them not to PM me asking for status updates. I want them to respect the 'Please don't bother asking about this' post in the forum.
05:46:48joshinLlorean: Tell me about it... Wife + 2 kids + work == not so much project time
05:47:31Lloreanhachi: I want them to not ask "does anyone even remember about that?" as if they think we've forgotten them.
05:51:07hachiwell, I'm sure that all us people are just wrong and assholes for ever asking, or being curious, or much of anything huh?
05:51:34LloreanI think that the phrasing asking if we've forgotten about you is very impolite.
05:51:42LloreanI think ignoring forum rules and starting a new thread on it is impolite.
05:52:01LloreanJust because someone's curious doesn't make the forum rules ignorable.
05:52:30LloreanDid you honestly think we'd forgotten about it, despite my editing the last post in the thread 3 weeks ago to further clarify?
05:53:00hachiI have no clue how to ask kindly though. the forum doesn't show for me that you edit it 3 weeks ago... last update I see is like 2 Feb
05:53:19LloreanIf you look at the last post, it says when it was last edited.
05:53:44LloreanI even explained why I locked the thread: Developers can post when there are updates, but we don't need people constantly asking the status.
05:54:29LloreanAnd you can ask kindly by saying "Hey guys, is there any progress on the 5.5G 80gb?" rather than phrasing it in a way where you've suggested that we've forgotten about it.
05:56:15LloreanMy problem isn't that you're asking.
05:56:28LloreanMy problem is how everyone seems to ask, choosing both impolite phrasing or ignoring posting rules.
05:56:58LloreanThere's nothing saying don't ask in here, just don't ask in the forums because it clutters the search, but you asked in a somewhat offensive way I thought.
05:57:13hachisorry bout that
05:57:23LloreanAnyway, status is this: Rockbox's FAT32 driver requires a lot more modification than the iPL one did because ours is very much trimmed down, and someone has to rewrite it.
05:57:49LloreanLinusN has suggested that he would do it, if/when he has time. Unless he finds time, or someone else volunteers to do it, that's where it'll stand.
05:58:58hachiI was ticked about the ipl fixes... I wanted to watch the progress on that as the patches and attempts were made... but the developers only made a monolithic patch and there was nothing to be learned about 'thought process' from it
05:59:29LloreanWell, we know what needs to be done.
05:59:40joshinIs there a key to the colors of the different patches in the tracker? I scanned the site and didn't see anything.
06:00:34Lloreanjoshin: Gray: Not yet modified/rated by a developer. Shades of Red reflect the severity from Low to Critical.
06:01:27hachiLlorean: I'm looking at it from this perspective, I'm an engineer... mostly perl, but I play in C too... I don't really have a clue about embedded system or driver programming... I wish I did because I'd love to poke at it to see if I can get anywhere with it... that's why I wanted to see thought process in a similar project
06:01:30Lloreanhachi: And sorry for jumping on you. People constantly PM me on the forums asking about that, or open new threads that I have to delete, so it's become a bit of a personal thorn in the side.
06:01:55joshinOk, thanks. Figured it was something along those lines. Now on to the documentation of how to build Rockbox from source and then to see if my brain can still think in C.
06:02:06n17ikh|Lappyoh, hey, there's something I've been meaning to ask about
06:02:30n17ikh|LappyI've got a few-days-old build on my iriver h10 20gb
06:02:50n17ikh|Lappyand when playing wavpack and being in the WPS screen, it pauses the music for a second or two about once every minute
06:02:55n17ikh|Lappyif you exit out of the WPS, it's fine
06:03:02Lloreann17ikh|Lappy: Does your WPS have peakmeters?
06:03:09n17ikh|Lappyer, one sec, I never looked
06:03:15n17ikh|Lappyor can't remember.
06:03:43Lloreann17ikh|Lappy: Peakmeters, Crossfeed, Equalizer, Replaygain, Dithering, all of these use CPU. wavpack is not the most CPU-efficient codec.
06:03:55LloreanThe WPS is more difficult to update (more CPU) than just filetree.
06:03:57n17ikh|LappyI am using the EQ
06:04:02LloreanEQ is one of the worst
06:04:12n17ikh|LappyI see.
06:04:22scorcheeach band adds to it...if you can get away with less bands, the better it is
06:04:49n17ikh|Lappyactually, I'm not using the EQ proper
06:04:56n17ikh|LappyI'm using the bass/treble adjustment though
06:05:28n17ikh|LappyI thought it might have been running out of buffer while spinning the hard drive up so I increased buffer time a little and it seemed to help
06:05:52n17ikh|Lappylike a combination of a slow to spin up disk and CPU troubles
06:05:53webguest89no, right? because there is no cpu scaling for the gigabeat
06:06:03webguest89hey i have a question... does the EQ, crossfeed, etc. impact the battery life of the gigabeat?
06:06:15Lloreanwebguest89: Yes.
06:06:20webguest89my two previous posts are inverted...
06:06:39Lloreann17ikh|Lappy: The anti-skip buffer can help a little, but it's really not meant for that.
06:06:46n17ikh|Lappyyeah, I know it's not
06:07:09n17ikh|Lappywebguest89: even though there's no cpu scaling the cpu at idle will use less power than a cpu under heavy load
06:07:12Lloreanwebguest89: The Gigabeat CPU draws less power if you use it less, so more CPU intensive actions do drain more battery
06:07:42webguest89then what's the purpose for cpu scaling?
06:08:01n17ikh|Lappybecause clocking down a very idle cpu will use much less power than leaving it at full speed
06:08:02LloreanOn the Gigabeat, not as much as elsewhere.
06:08:16n17ikh|Lappy(I'm thinking in normal computing terms here)
06:08:18LloreanIt doesn't seem to save power on the Gigabeat (and may even be worse)
06:08:36n17ikh|LappyI have no idea how the gigabeat's hardware actually handles idling and such
06:08:45n17ikh|Lappyso maybe I shouldn't answer questions I know nothing about >_>
06:09:05webguest89it seems like an inherent feature of any cpu so why is there a need to scale?
06:09:48webguest89thanks a bunch.
06:10:23scorchen17ikh|Lappy: i wish more people understood your last statement around here ;)
06:13:09*Llorean wonders how long the initial charge of a Gigabeat takes.
06:13:17scorcheSoap: SYN
06:17:20 Quit Yono ("Leaving")
06:17:22joshinWho's the best person to ping about wiki write accesses? I'm about to start the process of building RockBox for the Gigabeat and Sansa and figure that I can update the wiki as I go.
06:17:32joshinI did register my account.
06:18:21scorchethere isnt anything to update about the process of building
06:18:38scorchen17ikh|Lappy: you arent soap =/
06:19:20joshinHow about in "Find your target CPU" a listing of what CPU the Gigabeat needs. I'm not expecting to find anything major. Just nitpicky stuff that would get asked over and over here.
06:19:51*joshin needs to get to bed to reset his grammar processor...
06:20:18joshin"... a listing of what CPU target one should set for the Gigabeat F series"
06:20:24n17ikh|Lappyscorche: nope :/
06:20:41Lloreanjoshin: You don't set a CPU target.
06:20:50Lloreanjoshin: While compiling it asks what target player you're compiling for, and you answer.
06:21:20scorchethe only questions that get asked over and over are about not having the path set up and questions about updated to the wiki are really needed
06:21:36Lloreanscorche: Well, clarifications wouldn't be bad, really
06:21:46scorcheLlorean: such as?
06:21:49LloreanAs long as they aren't of the sort that show up in the SimpleGuideToCompiling.
06:22:06Lloreanscorche: I'm not sure, I'm just not willing to assume improvements can't be made.
06:22:48scorchetrue....i guess we will have to wait till after he is done then =)
06:23:49LloreanThe SimpleGuideToCompiling is anything but "simple" though, I think. It seems to talk in circles
06:24:25scorchei forgot who pointed it out, but the un-simple-version we have in the wiki is quite a bit shorter
06:25:13LloreanThe problem is ours assumes people can be told "you need to do this" rather than "type this" in a lot of cases.
06:25:18joshinLlorean: fine. I'm just at the reading of the wiki stage and looking to see if I can get a toolchain building as I sleep.
06:25:37Lloreanjoshin: You shouldn't need to do anything special to build it.
06:25:51scorchejoshin: have you looked at
06:25:58Lloreanjoshin: In Windows, you download one of our kits. Cygwin or VMWare or Colinux. On linux / OSX, you just run
06:26:09LloreanWhatever its name is. :-P
06:26:17scorcheLlorean: you had it down before!
06:26:30LloreanI've never been able to remember it.
06:26:33LloreanI just get lucky sometimes.
06:27:12joshinNice. I very much like how simple the OpenSlug folks made builds for the NSLU2. Then again, OpenEmbedded isn't that tough to deal with either. I'll grab and read
06:29:32 Quit jhulst ()
06:35:02 Join jhulst [0] (
06:36:19 Join ptw419 [0] (
06:37:49 Quit rotator ()
06:47:57 Join Br3nda [0] (
06:52:08 Join illriginal [0] (
06:54:06 Quit webguest89 ("CGI:IRC (EOF)")
06:59:10 Quit illriginal ("Leaving")
07:02:43 Quit RaRe ("The Evolution of Medicine - 2000 BC - Here, eat this root. 1000 AD - That root is heathen. Here, say this prayer. 1850 AD - T)
07:12:33***Saving seen data "./dancer.seen"
07:15:35 Quit markun (
07:15:35 Quit HEx (
07:15:35 Quit furiousD_ (
07:15:35 Quit dionoea (
07:15:35 Quit hachi (
07:16:02NJoinhachi [0] (
07:16:02NJoinmarkun [0] (n=markun@rockbox/developer/markun)
07:16:02NJoinHEx [0] (
07:16:02NJoinfuriousD_ [0] (
07:16:02NJoindionoea [0] (
07:16:44 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
07:31:39 Join davina [0] (
07:38:03*amiconn wonders about jhMikeS' commit msg
07:38:36 Join damaged_good [0] (
07:44:14pearldiverLlorean here?
07:44:44pearldiverdid you have a chance what version of the OF your F40 has?
07:44:52pearldiverto check*
07:45:28Lloreanpearldiver: Where do I check?
07:45:48pearldiverif i remember correctly its in the settings
07:46:00pearldiverversion info or something like that
07:46:35Lloreanpearldiver: 2.020US
07:47:25pearldiverme and toffe82 just agreed that all of his and mine gigabeats have serial numbers starting with a digit except for his rare'ish F10 which starts with X
07:47:59pearldiverso i thought theres something different on this Y
07:48:15pearldiverbut 2.02 is the common case with local ones
07:48:24LloreanBut the Champagne isn't supposed to be US release, is it?
07:48:48pearldiverwell you just confirmed that it was
07:49:22LloreanWith a weird serial number.
07:51:56LloreanThe system menu also says 4 played.
07:52:58toffe82you played 4 song with the original firmware :)
07:53:05LloreanExcept I didn't
07:53:17toffe82even if you upgrade, this number stays
07:53:23LloreanI went through the set time/date/timezone stuff, and went straight to the system menu
07:56:21pearldiversee the 1st one from the pictures?
07:56:34pearldiverthats the champaign i was hoping for
07:56:40Llorean Yeah, that's what I got
07:56:55LloreanUnder some light it looks just like a shinier silver, under other light you can see the difference
07:57:05pearldiverand even though it says champaign on the box it looks like the one towards the bottom
07:57:07 Part damaged_good
07:57:39LloreanThe 60?
07:57:50pearldivergigabeat F40(????(XS))
07:58:04pearldiverbefore the crazy colors start
07:58:09LloreanTry taking the champagne one out in direct sunlight and holding it at a slight angle
07:58:20LloreanIts tinting becomes much, much more apparent like that.
07:58:33pearldiverhmm i should try tomorrow
07:58:48LloreanI thought mine was silver for a while too, but finally got it under some very white lights.
07:58:51pearldiverbut i havent noticed that "champaign" tint under electric light that much
07:59:13LloreanI can't see it at all under any of the incandescents in my apartment
07:59:19LloreanBut in my kitchen, under flourescents, I can.
07:59:56scorcheLlorean: well, if you go outside, there is something called the sun.....almost makes going outside worth it ;)
08:00:16Lloreanscorche: I did suggest outside
08:00:23LloreanBut right now it's night here. :-
08:00:34scorcheLlorean: yeah, but would you deny me the chance to make a sub-par joke?
08:00:47scorches/joke/attempt at a joke
08:00:51 Join Jon-Kha [0] (
08:01:12Lloreanpearldiver: Oddly enough, through the lense of my digicam the coloring is MUCH more apparent.
08:01:30LloreanI was gonna take a comparison shot to show how subtle the difference is, but I don't think it'll work
08:04:20LloreanHow common are dead pixels on Gigabeats?
08:04:30pearldivernever been reported
08:05:18toffe82I have one
08:06:17LloreanI don't, on either, but I was wondering how lucky I am.
08:06:35 Join Rob2222 [0] (
08:07:03pearldiver1st time i see anyone mentions a dead pixel on gigabeat
08:07:28toffe82it is light green
08:08:13LloreanAlso, lightly shaking my gigabeat, there is a rattle. Common or no?
08:08:15pearldiveris it common among ipods?
08:08:19*amiconn didn't see dead pixels on lcds for a looong time
08:08:30pearldiverLlorean rattling is common
08:08:31Lloreanpearldiver: None of my MP3 players has one.
08:09:05 Join ender` [0] (i=krneki@
08:09:28LloreanBut every portable game system I've bought since the GBA has had one.
08:09:29toffe82on the 20 I worked on, I have only one with dead pixel
08:10:18 Join RaRe [0] (
08:10:47 Quit davina ("byeeeeee!")
08:12:14pearldiveri noticed for the last 5 days or so the battery life on my F20 is lower than usual
08:12:25pearldiveri need to do a bench
08:17:37toffe82good night
08:17:41 Part toffe82
08:21:48 Quit Br3nda (Remote closed the connection)
08:23:26 Quit Rob222241 (Read error: 110 (Connection timed out))
08:31:38 Join webguest69 [0] (i=c023110b@gateway/web/cgi-irc/
08:33:50 Join himitsu [0] (n=himitsu@
08:42:30 Join rift [0] (
08:44:44 Join RoC_MM [0] (
08:59:07 Quit RoC_MM ("Leaving")
09:01:29 Join spiorf [0] (
09:01:40 Quit JdGordon ("Konversation terminated!")
09:01:57 Quit webguest69 ("CGI:IRC")
09:09:59 Join petur [0] (i=d4efd6a6@gateway/web/cgi-irc/
09:10:20 Quit RaRe ("The Evolution of Medicine - 2000 BC - Here, eat this root. 1000 AD - That root is heathen. Here, say this prayer. 1850 AD - T)
09:12:34***Saving seen data "./dancer.seen"
09:13:51 Quit miepchen^schlaf (Read error: 60 (Operation timed out))
09:13:53 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
09:16:50 Join crop [0] (i=c27f0812@gateway/web/cgi-irc/
09:17:20cropHello. Assumed I have the lyrics id3 tag filled, how can I display the contents of the tag in a readable way?
09:19:20 Join JdGordon [0] (i=dced3920@gateway/web/cgi-irc/
09:21:20midkaycrop: from a coding standpoint, or..?
09:21:36midkaythere's no way to do it in rockbox currently, you'd have to code such an ability.
09:22:32markunthere is a plugin to display lyrics somewhere, you could take a look at that
09:22:58JdGordoncrop: thats another job for metadata on buffer...
09:23:33 Quit JdGordon (Client Quit)
09:23:55 Quit himitsu (Remote closed the connection)
09:24:06 Join Br3nda [0] (
09:27:42 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
09:27:50linuxstbamiconn: Am I right in thinking jhMikeS's SPDIF commit didn't actually change any code?
09:28:31LloreanIt certainly doesn't look like it did.
09:28:47cropmidkay: no, as a user.
09:28:53 Join himitsu [0] (n=himitsu@
09:29:01midkaycrop: no such way currently, then.
09:29:44crop midkay: :-( It's possible to show all mp3 tags of the song but I doubt that the lyrics will be readable then
09:30:10Lloreanlinuxstb: He did say he committed a bugfix in the forums, but that looks like it's just a clarification.
09:30:22cropI've searched the forums, there is a SNC plugin but it's not what I mean. My lyrics are in the tag, not as a separate file
09:30:32midkaycrop: not sure what you mean, "all mp3 tags".
09:30:38Lloreancrop: Well, lyrics aren't supported, as you were told.
09:30:53LloreanBut I'm pretty sure "View ID3 info" is still limited to the tags Rockbox supports.
09:30:58midkaycrop: well, as i said, there is no current way to do it. that's that. maybe at the end of the summer when metadata-on-buffer is (hopefully) done there will be a way.
09:31:29 Quit himitsu (Client Quit)
09:31:31midkayLlorean: i'm pretty sure of that as well. i can't see it magically supporting tags we haven't added support for..
09:31:37cropLlorean: Oh! I thought all id3 tags are displayed, i.e. all that are in the file
09:31:55cropmidkay: ok, I'll ive with that. Thanks for the info
09:32:21midkaycrop: no problem.
09:32:45linuxstbcrop: No, only the ones parsed and stored by Rockbox - there are many different id3v2 frames and Rockbox only handles some of them.
09:33:11cropI thought that the tags are also stored somewhere in a generic form, i.e. "key=value" and that only some tags are assigned to data structure fields and thus can be treated in a "special" way
09:33:34linuxstbid3v2 doesn't use key=value afaik.
09:33:36croplinuxstb: yes, now I see I was wrong
09:34:02 Join inversions [0] (
09:37:07LloreanBah, icons make me want to use non-tiny fonts.
09:37:09*Llorean mutters.
09:38:04LloreanI think I'm using a 9 or 10 tall font right now on my Gigabeat.
09:38:37JdGordonso it shows about 80 lines on the screen?
09:38:50midkayyou mean 32? :)
09:39:19LloreanIt shows "lots"
09:39:23LloreanWhich is good enough for me. :)
09:39:37dionoealinuxstb: is it expected that the dap (ipod video here) shuts down automatically when runing mpegplayer ?
09:39:49LloreanI'm slowly in the process of creating my own very, very tiny icon set.
09:39:58Lloreandionoea: It should not idle power off.
09:40:00dionoea(btw, looks like the 44.1kHz right, subs stayed synced for about 5 minutes ... before it shut down)
09:40:07dionoeahum ... weird
09:40:22linuxstbLlorean: Others have reported mpegplayer ignoring the idle poweroff.
09:40:30Lloreanlinuxstb: It never has for me.
09:40:36linuxstbEven on the Nano?
09:40:45LloreanI've watched a 22 minute video on the Nano
09:40:52linuxstbI can't remember it doing so either, but maybe I've disabled the idle poweroff....
09:40:59LloreanIt's enabled on my GB at least
09:41:15dionoeaI'll try disabling it in the code to see if it still shuts down
09:41:26LloreanIt'll idle power off if pause or in the menu
09:41:32LloreanBut not in videos, for me at least
09:41:40LloreanThough I've not tested much on my Nano.
09:42:37LloreanIt's set for 10 minutes on my Nano, and I got a full 22 last time I tried, but it was weeks ago
09:42:45markunmy gigabeat also never shutdown during video playback
09:43:41dionoeai guess that a bug in the code is highly unlikely to trigger a shutdown right? (It'd print some debug output on screen and freeze)
09:43:58linuxstbDoes it just power-off, or do a clean shutdown?
09:44:07linuxstbi.e. display the "Shutting down" splash?
09:44:23dionoeait displays a splash for like 1/2 a second
09:44:43dionoeai don't know what that was about ... but it did look like a power off
09:45:44markunlow power maybe?
09:45:49LloreanVery possible
09:46:09 Join barrywardell [0] (
09:46:14Llorean22 minutes of playback with backlight sapped a lot of battery on my Nano
09:46:46dionoeamarkun: well i could playback a full audio album afterwards (and battery level displayed 75%)
09:47:12dionoeaso low power would be kind of unexpected
09:47:52dionoeaunless the battery somehow freaks out if you draw too much power rapidly and then comes back to normal behavior when you restart
09:47:57Lloreandionoea: But you said it played for 5 minutes. Is your idle shutdown time reduced to 5 (the default is 10, I think)?
09:48:24 Quit Br3nda ("Konversation terminated!")
09:48:30dionoeawell maybe it was 10 :) (It was before the end of elephant dreams ... so less than 9 minutes right?)
09:49:20dionoeaI'll try to time it exactly next time
09:49:22LloreanYou could load elephants dream on a PC and find out how many minutes it was.
09:49:28LloreanJust fast forward to roughly where you were
09:49:33dionoeahehe :)
09:49:39LloreanIf it's some number like 7:30, then it's less likely that it was idle
09:50:06Llorean"Less" being the operative bit, since I don't know whether you adjusted volume 'n whatnot.
09:50:56LloreanBut that's weird if iPods are doing that.
09:55:09linuxstbLooking at the code, it looks like mpegplayer should idle-off...
09:55:48 Join norbusan [0] (
09:56:00 Part norbusan
09:59:19 Join himitsu [0] (n=himitsu@
10:02:12JdGordonamiconn: is the bitmap format variable always set to FORMAT_NATIVE when read with read_bmp_file()?
10:02:51cropRe displaying the lyrics: could we change the id3 parsing code so that not all known tags are stored in mp3_entry? Then a plugin could read them all in a generic way. And some (those supported by RB) would be stored in mp3_entry
10:04:24cropAs of now (iiuc), all registered tags are copied to mp3_struct and, possibly, postprocessed
10:04:24linuxstbcrop: There is very little room in an mp3_entry. We don't want to enlarge it.
10:05:10linuxstbBut that's what the "metadata-on-buffer" feature is intended to allow.
10:05:19croplinuxstb: I don't propose this. Just allow to register tags that won't be copied to mp3_struct. But allow to register a processing func for them.
10:05:55cropThen a plugin could read all of them as key-value pairs.
10:06:19cropAnd RB would only care about the supported tags.
10:06:41linuxstbWhere would the processing function live? And where would it store the data it's reading from the tags?
10:08:10croplinuxstb: the processing function could live anywhere. Some would in id3.c (as they are now), but others could be provided by the plugin. A flag should be added to the struct tag_resolver whether the tag is supported.
10:09:12cropOr we could have two funcs there: "main processing" and "postprocessing". "Main processing func" for the supported tags would just copy the tag to mp3_info.
10:09:35 Join roolku [0] (
10:09:42 Quit ptw419 ()
10:09:44cropAnd do nothing for unsupported tags
10:12:44 Quit inversions (Read error: 60 (Operation timed out))
10:12:48cropI.e. the parser could/should be given the array of tag_resolver. Internal RB parser would have its array (very similar to the current) but a plugin could provide other processing
10:13:24 Quit Llorean (Read error: 110 (Connection timed out))
10:15:31croplinuxstb: does this sound reasonable? Or do I misunderstand the RB way of parsing files?
10:15:33linuxstbcrop: If the parsing functions are living in core Rockbox, then why do we need a registration system? If they are in a plugin, then what happens when a user is running a different plugin when the playback engine is parsing the metadata for a new track?
10:15:43 Join bluebrother [0] (i=ZXmshA2l@rockbox/staff/bluebrother)
10:16:48croplinuxstb: the parsing func are passed as a parameter (fields in tag_resolver). So everyone can do with the tags what's needed.
10:17:09cropNo, not parsing func, but processing funcs.
10:17:30cropAnd also the set of the known tags.
10:17:54cropI.e. all that's in tag_resolver
10:18:39 Join brun0_ [0] (
10:20:16 Quit heanol_ (Remote closed the connection)
10:20:35cropA plugin would just reuse the available parsing code but would provide its own set of callbacks. The signature of the callbacks would need to change. It could be something like process_tag(char *tag_name, char *tag_value, struct mp3entry *entry)
10:21:39cropThe funcs provided by the plugin would e.g. just ignore the entry but rather collect the tags in a list of key-value pairs.
10:22:16cropAnd the funcs provided by the core RB would copy the supported tags to the entry
10:24:06 Join heanol [0] (
10:24:32 Join obo [0] (
10:25:39aliaskMy sim keeps crashing with *** glibc detected *** ./rockboxui: double free or corruption (fasttop): 0x0a26f688 *** when running a .rock I'm working on, would appreciate any help in the matter.
10:25:48croplinuxstb: no comments?
10:25:49linuxstbSo how big would the mp3_entry struct be?
10:27:10croplinuxstb: it wouldn't change at all
10:27:43linuxstbSo where would this extra data be stored?
10:27:44 Quit kkurbjun (Read error: 113 (No route to host))
10:27:57JdGordonaliask: gdb rockboxui
10:28:18croplinuxstb: in some buffer somewhere in the plugin.
10:28:26JdGordonaliask: your not trying to use threads or TSR plugin?
10:28:40linuxstbcrop: But what if the user is running a different plugin when the file needs to be parsed and buffered?
10:28:45aliaskJdGordon: Nope, just calling a plain ol' function.
10:28:55aliaskSeems to be where it crashes.
10:29:16JdGordongdb is your freind :p
10:29:28*aliask has never used it
10:29:45JdGordonrun gdb rockbox ui, then type r <enter>
10:30:01JdGordonwhen it crashes type bt <enter> and it will show you the callstack
10:30:09croplinuxstb: I think we are in different dimensions and you don't quite understand me. My proposal is that we decouple the parsing logic from the processing logic.
10:30:11JdGordonso you can see where it crashed and hopefully why
10:30:24 Quit rift (Read error: 104 (Connection reset by peer))
10:30:42 Quit atsea-87 (Remote closed the connection)
10:30:59linuxstbcrop: You're right, I don't understand... How can your plugin process something that isn't stored anywhere?
10:31:12croplinuxstb: the parser just finds the tag and calls the callback. Core RB provides its set of callbacks that store the supported tags in mp3_struct.
10:31:37linuxstbcrop: But you're saying the callback is in a plugin?
10:32:11cropBut a plugin (for example the (non existing) one for displaying lyrics) could provide another callback
10:32:42croplinuxstb: yes, the callback is in the caller of the parser. This can be core RB or a plugin
10:34:11linuxstbOK, so if the user wasn't running the plugin, the data would just be ignored?
10:34:34croplinuxstb: As of now, it's impossible to pass the set of the tags we're interested in and funcs for their processing from the outside. Everything is in id3.c. I'd like this to be changed.
10:35:17croplinuxstb: if the user wasn't running THAT plugin then yes, the data is ignored since the core RB doesn't care about it
10:36:03amiconnlinuxstb: You're right, jhMikeS' commit didn't change the actual code. It's just a more complicated way to write it
10:36:23*amiconn doesn't understand JdGordon's question
10:37:09JdGordonamiconn: the bmp loader, will it always set the format variable in the bitmap struct to native unless its a full colour bitmap?
10:37:20linuxstbcrop: OK, so you're proposing a mechanism for plugins to gain access to the track metadata?
10:37:59*amiconn still doesn't understand
10:38:12croplinuxstb: yes. And not necessarily the currently playing one.
10:38:20JdGordonamiconn: or... if you read a mono bitmap on a colour target, can you use that to be able to draw it with the mono funciton?
10:38:22amiconnThe bmp loader can be instructed to load a bmp either in native or mono format, or to auto-detect
10:38:44amiconnIn th elatter case, it will load 1-bit BMPs as mono, all others as native
10:38:45 Join atsea- [0] (i=ariel@gateway/tor/x-499c2aa9d59a23cf)
10:38:46linuxstbcrop: Go ahead and write a patch. Don't forget all the other tag formats Rockbox supports though.
10:38:54aliaskJdGordon: Thanks for pointing out gdb... pretty sure i've got it now
10:39:24croplinuxstb: you could select a file and "open with..." it with this plugin. And see all the tags in the file (and not only the supported)
10:39:38 Join RaRe [0] (
10:40:10JdGordonamiconn: int bmpformat = (FORMAT_NATIVE|FORMAT_DITHER); <- does that set it to only load it as native and not auto-detect?
10:40:23croplinuxstb: where does the parsing code reside? Only in id3.c?
10:41:06linuxstbcrop: For historical reasons (the hwcodec playback engine lives in firmware/) id3.c contains the id3 parsing code, and the rest (used by swcodec only) is in apps/metadata.c
10:41:32JdGordonthe wps code uses FORMAT_ANY|FORMAT_TRANSPARENT.. shoudl icons use that instead so we can load mono on colour targets?
10:41:49JdGordon... so they can be displayed with mono_bitmap_part
10:42:23croplinuxstb: ok. And we could onyl change this for id3 as a start. Hence the plugin would only work with mp3 files. And if that works good then this can be expanded to other formats as well.
10:43:14croplinuxstb: but I dont' have much time now so the plugin will not come very soon.
10:43:17amiconnFORMAT_ANY is for auto-detection, but then you need to check what you got from bmp_load
10:44:04 Quit barrywardell ()
10:44:06amiconnThat's why I said a few days ago you want to do the second step before the first. All this will become really simple when there will be only one bitmap drawing function instead of 3
10:45:14JdGordonhow far down the todo list is that?
10:47:06 Join Llorean [0] (n=Llorean@
10:50:55linuxstbJdGordon: Can you explain how the icon numbers in the .icons file and the viewers.config file are processed? There are still different icons listed in viewers.config for WAV files for example - so which icon is actually used?
10:51:20JdGordonI thought I changed the wav icons so they were both the same?
10:51:41linuxstbThere are 4 entries for WAV, two with "-" and two with "10".
10:51:53JdGordonok, i'm going to try explainig it on the wiki better, Ive explaired it 3 times in irc already :p
10:52:19linuxstbAlso, txt and jpg have the same icon...
10:52:23JdGordonwhich part dont you get? the numbers? or which icon is used?
10:52:47linuxstbHow the various files are processed to decide what icon is used for a particular extension.
10:53:02JdGordonok, 5 min
10:53:45 Join Siltaar [0] (
10:54:11linuxstbIn fact, txt is listed three times in viewers.config, with three different icon numbers...
10:55:05linuxstbviewers.config isn't mentioned at all on the CustomIcons wiki page.
10:55:42Lloreanviewers.config shouldn't be modified.
10:55:50LloreanThe .icons file is basically an override for it.
10:56:19linuxstbWhy bother with that? Why not just always have a .icons file and leave icons out of viewers.config?
10:56:21LloreanThat way people aren't including replacement viewers.config files with themes that overwrite the one in the .rockbox folder and remove any custom lines.
10:56:32Lloreanlinuxstb: That's what I suggested, I'm not sure why it wasn't done that way.
10:56:59 Join Entasis [0] (
10:59:36JdGordonpossibly one less file to parse if it stays in the .config
10:59:49JdGordon1 less to make sure gets into the build..
11:02:27JdGordonhopefully that clears it up a bit
11:04:05 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
11:04:17linuxstbSo you could just remove the icon field from viewers.config for all already-supported filetypes?
11:05:03JdGordonexcept, aparently wav needs an icon for archos?
11:05:04 Join miepchen^schlaf [0] (
11:06:10amiconnlinuxstb: Without .icons, viewers.config decides which icon number is used for viewer-supporte3d formats
11:06:34amiconnI don't think that should be removed, unless we aways want to ship an .icons file
11:06:46JdGordonwell... its more the other way around... .icons overrides .config..
11:07:23*amiconn wonders how this is the other way round
11:07:47JdGordonumm... hmm....
11:07:49*JdGordon shuts up
11:08:18JdGordonyour way sounds like the .icons is checked for before viewers.config...
11:08:40 Join rift_ [0] (
11:08:45 Quit pearldiver (Read error: 110 (Connection timed out))
11:08:54 Quit crop ("CGI:IRC")
11:10:16 Join barrywardell [0] (
11:12:36 Join Br3nda [0] (
11:12:39***Saving seen data "./dancer.seen"
11:13:43linuxstbSo what's involved in a user adding a new file extension to viewers.config? They'll need to refer to an icon in the viewers icon file, which may be in a different order depending on the iconset author?
11:13:49 Join PaulJam [0] (
11:14:27JdGordonminimum they do is add the extension and a plugin
11:14:49linuxstbBut if they want an icon to appear?
11:14:49JdGordonwhich has always been the case
11:15:27JdGordonthen they would add the icon to the default viewers.bmp for all targets
11:16:22*JdGordon see's about getting some tucker, back later
11:20:35PaulJamJdGordon: i noticed, that Icon_Cursor and Icon_Reverse_Cursor are used in the virtual keyboard, which uses the systemfont (unless you use a custom layout) and this looks a little bit weird with icons that are larger than the default ones.
11:22:58 Quit juxtap (Read error: 60 (Operation timed out))
11:26:58spiorfi think i haven't understood this thing: it's viewers.config or theme.icons that decides the icons to use?
11:27:18Lloreanspiorf: viewers.config unless there's a theme.icons
11:27:43Lloreantheme.icons is an override file if they want to put their viewer icons in a different order, or have more than the standard viewer bitmap.
11:32:24PaulJambtw, are the icons 8 (magnifying glass) and 9 (double note) in the viewers.bmp actually used somewhere in rockbox?
11:35:06 Join pondlife [0] (
11:37:53linuxstbamiconn: I'm talking about removing any numbers from viewers.config which aren't actually used. IIUC, if when a viewers.config is processed, if there is already a built-in icon for that extension, then the icon number in viewers.config is ignored.
11:38:15amiconnIs it?
11:38:22linuxstbamiconn: So for WAV, we would need to keep the icon for the wavplay entry, but not the others.
11:38:45amiconnAfaik, before themable icons, no viewers-supported extension used a built-in icon
11:38:49linuxstbThat's what JdGordon just wrote on the wiki - "this means that icons for already known filetypes are ignored from this file." (regarding viewers.config)
11:39:20Lloreanamiconn: mp3encode and a few others did, at times, I believe.
11:39:24Lloreanamiconn: The wav viewer...
11:40:04amiconnBut then, the icon in viewers.config was ignored before (set to 00 00 00 00 00 00)
11:40:06 Join Llorea1 [0] (
11:40:21 Quit Llorean (Nick collision from services.)
11:40:25 Nick Llorea1 is now known as Llorean (
11:40:26amiconnSame if multiple viewers are associated with the same extension
11:41:58LloreanSo why not remove any duplicates, and any that have builtin icons?
11:44:17amiconnRemove from where?
11:44:58linuxstbI'm talking about removing any icon entries from viewers.config which aren't actually being processed.
11:45:14linuxstbJust to make it clearer what's going on.
11:45:26 Quit atsea- (Read error: 104 (Connection reset by peer))
11:45:47Lloreanamiconn: In viewers.config, remove the icon numbers.
11:45:54LloreanFor the ones that aren't actually used.
11:46:54PaulJamLlorean: i could imagine that removing duplicates from the viewes.config could confuse themers, because if one line for txt files has a "1" at the end and one a "-" it would be unclear if the icon is inbuilt or taken from the viewers.bmp
11:47:43LloreanPaulJam: Make the first one the numbered one, all latter ones -, then I think the assumption will be "It's already set, up above, no need to re-set"
11:48:16linuxstbPaulJam: That's my point - why does viewers.config contain 3 different icons for .txt ?
11:48:46PaulJamand extensions that use inbuilt icons already have "-" at the end of the line (except wav and bmp, but i guess these are special cases)
11:50:18amiconnlinuxstb: It doesn't. I can only see 2 (the second one most likely being a mistake). The third is '-', meaning "already set"
11:50:41linuxstbamiconn: For which extension?
11:51:33 Quit Br3nda ("Konversation terminated!")
11:52:50linuxstbThe three entries I see are "1", "2" (meaning take the icon from the viewers icon.bmp) and "-" meaning "no icon".
11:53:29linuxstbThe wiki page defines "-" as meaning "no icon".
11:53:35linuxstbAt least in the "icons" file...
11:55:08amiconnThe 2 is a mistake, should also be -
11:55:41amiconnObviously the same extension cannot have multiple icons
11:56:09PaulJamin my opinion it would be clearer if they were all set to "1"
12:01:16amiconnBut it would be wrong - the entry isn't used
12:01:36linuxstbI'm thinking it could simplify things if all extensions listed in the default viewers.config had icons in the main iconset. That would leave the second icon file purely for user-added extensions (so wouldn't be needed for most users) and icon information could be removed from viewers.config completely.
12:02:37amiconnviewers.config exists because we do not want the list of viewer-supported files built in
12:03:01amiconnMaking the icons for them built-in would break that concept
12:03:09LloreanOut of curiosity, why don't we want that list built in?
12:03:35LloreanAt least, the default list that is.
12:03:49linuxstbI think of the viewers the same as the codecs - they're only external to cut down on binary size.
12:04:25Lloreanlinuxstb: The problem with viewers is that you may want to associate them with many more filetypes
12:04:25amiconnBecause they are dynamic. E.g. if someone doesn't need support for jpeg, he could take out the appropriate line from viewers.config and delete the viewer
12:04:25linuxstbthird-party viewers and codecs just haven't happened...
12:04:25LloreanFor example, .1st, .nfo, .html, etc could all validly be associated with viewer
12:04:55amiconnWith the viewers.config system, jpg files will then no longer be displayed as supported
12:04:59LloreanI still think the best solution is to just include a default .icons file to go with the default viewer icons.
12:05:05LloreanAnd strip icons from viewers.config entirely
12:05:24linuxstbamiconn: Yes, and they could still do that. If no viewers are mapped to an extension (and that extension isn't used by the core), then they wouldn't be shown as supported.
12:05:32amiconnlinuxstb: Following your binary size argument, making the viewers supported icons bult-in doesn't make sense...
12:05:58linuxstbI thought the icons were now loaded from an external .bmp?
12:06:24linuxstbBut yes, there's still the increase in size from hard-coding the supported extension.
12:06:44amiconnThe default icons for core-supported formats are bult-in as before
12:07:13 Join mots [0] (n=mots@
12:07:20amiconnThe default viewer icons are external as before, just not defined directly in viewers.config but as a .bmp referenced by viewers.config
12:07:53motsis it normal that the rockboxpatcher for ipods takes years for "scanning disk devices"?
12:08:06linuxstbamiconn: OK, I didn't realise that...
12:08:07amiconnAnd the viewers-supported extensions weren't hardcoded before and aren't now either.
12:08:07GodEatermots: what's rockboxpatcher ?
12:08:47linuxstbmots: It was only released a few months ago...
12:09:08GodEaterlinuxstb: hehe - good answer
12:09:34roolkuamiconn: why do icons need to be related to viewers (or even core support) anyway? they could be completely disassociated, just have a built in extension->icon mapping which can be overridden by a theme set
12:09:50linuxstbmots: It should only take as long as any disks in your system which may be sleeping take to wake up. So if it's more than about a minute, something is wrong...
12:09:53roolkuseems the least complex and most flexible approach
12:10:36Lloreanroolku: If they were removed from the viewers.config file, and always used a .icons file, that would essentially be the case.
12:15:34 Quit scorche (Read error: 104 (Connection reset by peer))
12:19:07rift_i tried to create my own icon theme
12:19:19rift_with 11x12icones
12:19:38 Join scorche` [0] (i=scorche@rockbox/administrator/scorche)
12:19:49rift_but rockbox doesn't "cut" perfectly the icone from de toolbar
12:19:53 Nick scorche` is now known as scorche (i=scorche@rockbox/administrator/scorche)
12:20:04 Join Thundercloud [0] (
12:20:44rift_maybe i have to specify hte size of my icones ?
12:21:02PaulJamrift_: i'm not sure what you mean
12:21:54Lloreanrift_: I believe Rockbox determines your icon size by your main icon bitmap.
12:22:07LloreanHow many pixels tall, total, is it?
12:22:34LloreanThat's not the right number
12:22:44LloreanThat would give an icon height of 11.625
12:23:02PaulJamit should be 12x32 = 384
12:23:05LloreanThere are supposed to be 32 main icons, so if your icons are N tall, the image should be N*32
12:23:10rift_i have forgotten one icon
12:27:32 Quit miepchen^schlaf ("Verlassend")
12:27:57Lloreanpondlife: Oops. I read the post yesterday, and then responded to it today without re-reading it. =/
12:29:41 Quit mots (Read error: 104 (Connection reset by peer))
12:31:34 Join miepchen^schlaf [0] (
12:31:52 Join Br3nda [0] (
12:31:53pondlifeLlorean: We do have too many options
12:32:27LloreanPerhaps, but those ones are actually useful.
12:32:50 Quit Siltaar (Remote closed the connection)
12:32:51pondlifeI guess you might want to use the hold switch to force the backlight on?
12:32:52PaulJamhmm, there seems to be something wrong with the setting for the line selector. if i set it to pointer it doesn't write anything to the config.cfg and when i set it to Bar (inverse) it writes "invert cursor: off" to the config.cfg
12:33:00Lloreanpondlife: I force it off, rather than on
12:33:05pondlifeAh, ok.
12:33:37Lloreanpondlife: I have backlight set to a long timeout (20 second) with hold turning it off. That way it rarely fades on me while I play Bejeweled or something similar (anything I stop to think in), but it will fade if I leave it without remembering to turn on hold
12:33:41PaulJamI'm using the hold switch to torn off the backlight too.
12:33:42LloreanMeanwhile hold lets me force it off quickly
12:34:27 Join Siltaar [0] (
12:35:18pondlifeWow, amazon are doing a cheap-ish Sansa too..
12:35:37 Quit lini (Read error: 110 (Connection timed out))
12:37:21Lloreanpondlife: I think I can get it for $20 US less if I just go down the street to best buy
12:37:49pondlifeBut that's cheap for the UK 80GB model.
12:37:57LloreanAh, perhaps in the UK, yes.
12:38:03Llorean8gb, by the way
12:38:34pondlifeAh I saw e280.... my mistake
12:38:42pondlifeNot so good then!
12:38:53LloreanThe sansas only come in 2, 4, 6, and 8
12:39:07LloreanAnd that price is pretty much standard for the 8 here in the US. Or at least in the standard range.
12:39:16LloreanI wish they'd used sensible numbers
12:39:17pondlifeGuess I'll wait for those 80GB Toshiba drives to get cheaper.
12:39:37LloreanWhy the 250 = 2, 260 = 4, 270 = 6, I'll never guess
12:39:48LloreanThe 280 = 8 was just luck on their part when they decided to release a new model later
12:39:49pondlifeTo confuse me I guess
12:40:37linuxstbSupply still seems very restricted on those 80GB drives - maybe I didn't look well enough, but I could only find one seller selling them.
12:40:49pondlifeIndeed, and that's not helping prices
12:41:02amiconnroolku: The point is that the extension->icon mapping for extensions supported by viewers should *not* be built in
12:41:50amiconnRight now viewers.config defines both the icon and the viewer associated with an extension, which is the right way imho
12:42:06Lloreanamiconn: I agree it shouldn't be built in, but why not split the extension->viewer mapping from the extension->icon mapping?
12:42:07amiconnThe only deviation is that only the first icon mapping matters
12:42:14LloreanWhy is the right way to define both the icon and the viewer in the same file?
12:42:20LloreanThe icon represents the extension, not the viewer.
12:42:37amiconnThat's true, but then we need to load 2 files instead of one
12:42:42rift_ i tried to compile rockbox simulator for ipod nano on freebsd, but y have an error,
12:42:45rift_gmake[1]: *** No rule to make target `SDL.h', needed by `/usr/home/rift/src/rockbox/sim/sim/button.o'. Stop.
12:43:41Lloreanamiconn: Is that really so bad, to load 2 files instead of 1, if it provides a clearer separation of concepts?
12:44:21pondlifeLlorean: I'm amazed you don't get tired of moving forum posts; it seems rare that anyone actually starts a topic in the right place :(
12:45:09Lloreanpondlife: Yeah, I need to figure out how to make the naming clearer I think. I blame myself. I have an intent, but I've not communicated it as clearly as I'd like, I guess.
12:45:15pondlifeThe installation/removal sections are much abused.
12:45:33LloreanIt's mostly the player-specific areas, because people think of them as everything player-specific rather than "Installation/Removal questions"
12:45:45pondlifeMaybe they should all be merged? People see their player type and go straight in.
12:45:57pondlifei.e. have a Getting Started section instead?
12:46:38pondlife"ooh ooh, I've got one of those!"
12:50:51LloreanWell, the installation process is the only really... unique-by-player stuff. Or at least, it's 90% of the the uniqueness of any given player
12:51:00JdGordonPaulJam: that was part of my sinister plan to get someone to completly redo the vkeyboard!
12:51:03LloreanI was tempted to break it up by install process instead.
12:51:45LloreanLike, have a section named "FWPatcher related installs" with the subtitle "H100, H300 installation help" and "Ipodpatcher Installs" subtitled "Help for installing on all iPods" and so on.
12:55:09LloreanIf RBUtil gets finished, they can all be replaced with an "Installation Help" forum, probably.
12:56:51XavierGrnow that the UI stasrts to become very beautifull I would like to see a modified status bar, at least the font size seems to be critical for targets with big screens
12:57:20amiconnpetur, pondlife: One of you wanted proper unicode support in the windows sim. Wanna try a patch?
12:57:28JdGordonthats pretty much the last thing that needs prety-ifiyng///
12:57:32pondlifeYes please
12:57:47LloreanThe only characters the status bar ever uses are 0-9, :, and %, right?
12:58:17JdGordonLlorean: the rec screen does some funny stukk with it
12:58:19XavierGrand the volume and battery graph, but yeah these are not characters
12:58:20JdGordonbut usually, yes
12:58:29LloreanJdGordon: Then the status bar could be represented by a single bitmap.
12:58:38JdGordonI think the status bar should go the wps route
12:59:34XavierGrJdGordon: what about the status bar on the menu? or you mean to have parsing for the status bar like wps?
13:00:16amiconnpondlife, petur:">
13:00:19LloreanThe status bar could have its own limited font buffer, for just those 12 characters, so that it could use any font regardless of UI font?
13:00:55JdGordonXavierGr: I think we wait till viewports, then make the statusbar a seperate vieport which can be customised like the wps
13:01:00LloreanThen the other symbols it used could be defined in a .bmp, and the status bar definition file could just let you pick the order each status bar widget is displayed from left to right, and spacing between 'em
13:01:11LloreanI don't think it needs strong customiziation.
13:01:38LloreanJust, using pseudotags, [battery meter][13 pixels][volume meter][12 pixels][clock]
13:01:38amiconnLlorean: The battery and volume graph are calculated
13:01:42JdGordonit would be nice to have the list title in the status bar if font/lcd allowed it
13:01:52Lloreanamiconn: They can be with .bmps too
13:02:05amiconnEspecially battery can't easily be composed with bitmaps
13:02:09Lloreanamiconn: Battery frame, always drawn in full, battery filler, drawn horizontally L->R based on %
13:02:18LloreanMuch like progress bars in WPS
13:02:22amiconnThere are 2 fillers: black and grey
13:02:28amiconn(execpt on mono lcds)
13:02:28linuxstbLlorean: What about the charging animation?
13:02:33Lloreanamiconn: Then two bitmaps
13:02:43Lloreanlinuxstb: So two filler images instead of one.
13:04:14roolkuamiconn: Your argument is only valid if you assume an association between viewers and icons. I want to remove that association. Icons should be mappable to extension whether there is one, more or now viewer available for a particular association (and also available viewers don't mean there has to be an icon)
13:04:45Lloreanroolku: You can currently cover all of those situations, it's just not apparent that you can
13:04:49JdGordonfor simplicity icons are associated with extensions, not plugins
13:04:50roolkubut I can see people might be insistent on the viewer to icon association
13:06:15roolkucurrently the viewers to extension relationship is a many-to-many which causes all these problems
13:06:41roolkua 1-to-1 relationship between extensions and icons would be a lot more KISS
13:06:55amiconnroolku: My argument wasn't about the exact association. Even if you make icon association independent of viewer association, we still don't want the extension->icon association to be built-in except for core-supported formats
13:08:08 Quit darkless ("Leaving")
13:09:42 Join Nico_P [0] (
13:10:09JdGordonhey Nico_P
13:10:16JdGordonso whats the story with the text height patch?
13:10:36linuxstbNico_P: And how's MoB? :)
13:10:37Nico_Pit's still on th tracker and I intend on finishing and committing it
13:10:59Nico_Plinuxstb: MoB will soon become my main interest :)
13:11:11LloreanNico_P: It had better! :-P
13:11:22Nico_PJdGordon: it's almost finished but I need to add the tag to a bunch of WPSs
13:11:31JdGordonok :)
13:11:54Nico_PJdGordon: but I don't have much time for rockbox this week, my gf is with me in toulouse
13:12:04Nico_PLlorean: :)
13:12:18XavierGrI wonder why the gigabeat wants you to press the power button ages to start up
13:12:26JdGordonNico_P: you dont have to make up excuses for us :D
13:12:43***Saving seen data "./dancer.seen"
13:12:45LloreanXavierGr: A great, great fear of accidental power on?
13:13:11XavierGrLlorean: heh, it isn't a self destruct button anyway :P
13:13:17 Join matsl [0] (
13:13:29LloreanXavierGr: Haven't you been reading about those exploding batteries in the news?
13:13:37XavierGrbattery bench for F40 with 128-320 mp3s = 16 hours and 40 minutes
13:13:52pondlifeamiconn: I need Unifont for this, right?
13:13:53XavierGris that good?
13:14:04LloreanXavierGr: A little above average, I think
13:14:16amiconnpondlife: Well, if you want to test exotic scripts, then yes
13:14:31amiconnBut even german umlauts or accented letters didn't work before
13:14:45XavierGrLlorean: I thought that the average was about 20 hours
13:14:51pondlifeOK, just checking with the umlauts in Bjork..
13:15:24XavierGrLlorean: Yeah I am just going to upload my score there
13:15:30linuxstbNico_P: BTW, why do you need to add the line-height tag to WPSs? Is it a compulsory tag?
13:15:36LloreanThere is one 19h one, but most of them are slightly under 16
13:16:09Lloreanlinuxstb: Some of the built-in WPSes are compatible with it though (rockboxed, icatcher)
13:16:35peturamiconn: sorry no time now
13:17:20pondlifeamiconn: Well, it passes the Bjork test.
13:17:28pondlifeWill carry on playing with it!
13:18:19linuxstbLlorean: OK, then I misunderstood Nico_P's use of the phrase "need to"... Adding the line-height tag to WPSs isn't needed...
13:19:55LloreanYeah, his personal desire makes him need to, they don't require it. :)
13:20:03LloreanAt least, I don't think they do...
13:21:28PaulJambtw, can a wps then contain multiple line height tags or just one?
13:21:58amiconnpondlife: I'm not fully satisfied with how the patch works, but at least it doesn't add a ton of ifdefs
13:22:23pondlifeamiconn: I can crash using the patch and JPG viewer.
13:22:45pondlifeJust checking I did a make install properly....
13:22:54LloreanPaulJam: It's just a global line height for the WPS.
13:23:05amiconnIt adds quite a number of macros though, and those utf8<->ucs2 conversion functions with static buffers
13:23:07PaulJamok, thanks
13:23:17LloreanPaulJam: The screen is still divided into a number of equal height lines, it just removes the link between the height of the lines and the height of the font.
13:23:56*amiconn also wants to make the console show unicode chars properly
13:24:10amiconnI recently learned that the windows console *does* support utf-8
13:24:20amiconnIt's code page 65001
13:24:58roolkusorry, work interfered
13:25:30roolkuamiconn: what is the issue with built-in ? It can be very minimalistic
13:26:09roolkuI see it the same as the default wps
13:26:28amiconn(1) It takes binary space. (2) It's static. Currently you can make a file type unsupported by removing it from viewers.config
13:26:28markunamiconn: good to know!
13:27:02amiconnmarkun: Yeah, just I need to find out how to set this code page for a console application
13:27:16linuxstbamiconn: Have a built-in set of icons doesn't affect 2) - that can still be based on viewers.config.
13:27:18roolkuamiconn: 1) more than currently? why?
13:27:19pondlifeamiconn: Try playing a track with an umlaut in the pathname, then go back to the file viewer and select a JPG
13:27:20JdGordonamiconn: well, in that case the whole table from tree.c should be removed
13:27:47JdGordon... which it could be btw... fairly easily
13:28:02linuxstbroolku: The built-in filetypes have default icons in the binary.
13:28:02roolkuamiconn: 2) what do you mean by "unsupported" I suggested to dissassocuate icons from viewers
13:29:15pondlifeamiconn: gdb only gives a little info - Segmentation fault at SDL_sysmutex.c:90
13:29:34pondlifeAnd no useful backtrace
13:29:35amiconnWell, true. But unsupported extensions shouldn't have an icon either. Or at least it should be possible to choose whether I want one
13:29:57JdGordonthe file type table from tree.c uses about 640bytes of bin space which could be loaded into RAM on boot from a txt file
13:29:59roolkulinuxstb: yes? I can't follow the conclusion.
13:30:14JdGordonamiconn: unsupported files always get no icon
13:30:14 Quit miepchen^schlaf (Read error: 113 (No route to host))
13:30:29pondlifeJdGordon: I thought only filetypes.c was meant to have extension-specific code...
13:30:32LloreanJdGordon: I thought you could define an icon for *any* filetype in the .icons file
13:30:35amiconnJdGordon: Not if the list is built-in as roolku suggested
13:30:42roolkuthe default could be no icons for any files that have a viewer associated with them
13:31:19JdGordonpondlife: thats what i was complainig about yestersday... filetype icons are handled in 3 files...
13:31:23JdGordonLlorean: no, only supported files
13:31:30 Join atsea- [0] (i=ariel@gateway/tor/x-a7aae23b6f50b757)
13:31:42JdGordonamiconn: no, i;m agreeing with you, that it could (should?) be completly moved out of the bin
13:31:56LloreanJdGordon: In my mind the .icons file should allow you to set associations between any filetype and icon, whether there's a viewer for them or not.
13:32:01roolkuif we ship an default mapping file (including bmp) as a starting point this can be used for the theme builders
13:32:23amiconnJdGordon: It should neither be completely built-in nor completely moved out
13:32:33roolkucompletely moved out :)
13:32:51JdGordonLlorean: tough :) its impossible to do atm... it does string comparisons on the extensions which are stored in the audio buffer, so adding after boot is painful
13:33:08amiconnThe point is, core supported formats should work and have their icon without needing any extzernal files
13:33:36amiconn(of course it might be desirable to ovverride built-in icons for those who care that much about lloks)
13:34:00roolkuthe built-in I only suggested as we also have a default wps built-in
13:34:16JdGordonwasnt it decided though, that without a full zip install rockbox was useless anyway? so really no files would be supproted?
13:34:37JdGordonso it would be ok to move even the inbuilt filetypes out of the bin into a txt file
13:34:57roolkuwould get my vote
13:35:28*JdGordon not even sure why that table is in tree.c, a quick search doesnt show any reason for it to be there and not filetypes.c
13:35:49LloreanWhy exactly is moving the inbuilt filetypes out a good idea again?
13:35:52LloreanWhat benefit does it gain?
13:35:54roolkuone single list that covers all extensions
13:36:12JdGordonsmall binary decrease also
13:36:22amiconnroolku: Yes, as wps is a core-supported format
13:36:28roolkufor paranoid people a fall-back could be compiled in (but I see this as optional)
13:36:28JdGordonassuming the text parsing functino would use less space than the table
13:37:12amiconnLlorean: You can't simply rely on an external file, you need a fallback in case this file doesn't exists, is unreadable, garbled etc
13:37:25*JdGordon disagrees
13:37:27Lloreanamiconn: I'm for keeping them inbuilt.
13:37:30pondlifeamiconn: Did you see my patch bug report?
13:37:32roolkuamiconn: I didn't mean a wps icon - I mean a fall back if no external wps is loaded
13:38:14roolkuamiconn: so we could have a fall back for if no external extension-icon file is loaded
13:38:20amiconnJdGordon: Rockbox might be unusable with no accompanying files on swcodec because you simply need the codecs for playing music. This isn't true on hwcodec
13:38:29roolkuwhich could be very minimalistic (to save space in binary)
13:38:34 Quit Br3nda ("Konversation terminated!")
13:38:51JdGordonamiconn: ah, yes, forgot about hwcodec...
13:38:51roolkuand for example only include core-supported extensions
13:39:46*roolku doesn't consider icons important functionality
13:39:56roolkuconsidering they can be turned off
13:40:17pondlifeIt can fall back to no or blank icons easily enough
13:40:25roolkuso the HWCODEC would still work without an external icon file
13:41:08JdGordonroolku: the icons are only coincidental to the discussion... the inbuilt icons use the same number for their icon as the tree uses for its attributes
13:41:56linuxstbJdGordon: I thought the discussion was about icons...
13:42:04preglowbtw, am i the only one not quite seeing the point with icons in menus?
13:42:10roolkuJdGordon: I don't know what you mean?
13:42:23preglowlike one for every menu entry
13:42:39peturpreglow: eye-candy
13:42:53roolkupreglow: it looks nice on targets with high resolution, but I turn them off on lowres screens like h140
13:42:56preglowcluttering eye-candy :/
13:43:02JdGordonoh woops.... no ignore me
13:43:17preglowi'll agree they work better on hires screens, yes
13:43:19JdGordon4 bytes per built in is used for the icon
13:43:27 Quit elinenbe_ (Read error: 104 (Connection reset by peer))
13:43:30preglowbut i still think they're just cluttering the screen instead of making it prettier
13:44:41markunanybody against moving all the mi4 based portalplayer devices to firmware/arm/target/portalplayer ?
13:44:43LloreanSo, Markun mentioned something to me, and now I'm curious. Is there a reason that under architecture (say, /arm/) in target tree, we have it split only by manufacturer rather than categories like /mi4/ or /portalplayer/ since those share more similarities based on their standard design than their manufacturer description?
13:45:26preglowdunno, i think linusn was the one who did the initial directory layout
13:45:47 Join webguest06 [0] (i=c023110b@gateway/web/cgi-irc/
13:45:58markunalso, the Gigabeat F, X and Kenwood HD20GA7, HD30GA9 are built on a platform codenamed "Legna" by Toshiba
13:46:01LinusNwhen i did that layout, there was only one portalplayer target
13:46:06markunthe Gigabeat S, V and Zune are called "Ottoman"
13:46:14markundo you think we could use those names in the target tree?
13:46:18LloreanLinusN: Well the real question is: Any objections to changing it, as it seems to make sense to?
13:46:28markunfor example
13:46:30LloreanAdd "platforms" rather than manufacturers?
13:46:34JdGordondoesnt the path use those dirs for #include search?
13:46:46LinusNJdGordon: yes
13:46:47JdGordonso we can only go 3 deep without changing more scripts?
13:46:59LloreanJdGordon: In those cases it would be replacing the manufacturer level
13:47:03LloreanIt's not adding a new level.
13:47:18LinusNbut yes, we could call it "platform" instead
13:47:22LloreanWe still have say /iRiver/ for the iFP, since it doesn't have an abstract platform it falls into.
13:47:34markunLlorean: agreed
13:47:54linuxstbThat could be the CPU in the iFP. ARM is a special case because of all these variants...
13:47:58LloreanBut the Gigabeat F and X go on /Legna/ since that's the reference design / platform / whatever we call it.
13:48:15JdGordonmarkun: did you mean firmware/target/arm/pp ? not firmware/arm/target/pp?
13:48:29markunJdGordon: of course :)
13:48:37JdGordonok, then it makes sense
13:49:02JdGordonand shouldnt /ipod be /apple?
13:49:33LloreanJdGordon: I think /iPod/ might be more descriptive as a platform
13:49:56LloreanBut, shrug.
13:50:22LloreanThe real fix is for the million potential MI4 targets that could end up showing up, plus the Toshiba families.
13:50:40JdGordonand move all those .c's out of /arm ?
13:50:52markunJdGordon: yes, I would say so
13:51:09JdGordongo for it then :)
13:52:04markunso, will it be "pp", "portalplayer" or "mi4"?
13:52:24markunor something else :)
13:52:31markunA good thing we have svn now..
13:53:02markunJdGordon: these player probably share quite some code as well:
13:53:17markunif we ever support them
13:53:26linuxstbmarkun: I think I'm missing something - the ipods will also be under the new portalplayer directory?
13:53:44markunlinuxstb: I don't know
13:54:39markunor do we have to think about a non hyrarchical way to organize the code?
13:54:48linuxstbWe could just have arm/cputype/device/
13:55:22markunmaybe better even
13:55:29linuxstbAnd get rid of manufacturer completely for the arm devices, as they seem to be mostly SoCs
13:55:37markuns3c2440, portalplayer, pnx0101
13:55:53markunmaybe soon imx31
13:56:18markunLinusN: what do you think?
13:56:30JdGordonlinuxstb: so cputype would be pp50** or?
13:58:03linuxstbI don't think we need to split the portalplayers. Looking at ata-pp5002.c and ata-pp5020.c, they could be unified by using #defines - not even any #ifdef are needed.
13:58:44linuxstbThey seem the only files split by PP chip.
13:59:30LloreanSo a folder for the PP 50 series?
14:00:43linuxstbAre there any other series?
14:00:55LloreanNot that we've encountered yet, at least
14:01:02JdGordon /arm/pp/{ipod{-3g,-video}, sansa-e200, iriver-h10} ? wont that be more messy than what we have already?
14:01:04linuxstbAnd as the company is dead...
14:01:10JdGordonunless im completly mistaken
14:01:23amiconnlinuxstb: yet. What about iram size?
14:01:26 Join juxtap [0] (
14:01:40linuxstbamiconn: What do you mean?
14:01:54amiconnDifferences between pp50* chips...
14:02:57linuxstbIsn't that just a #define? That should require different directories in target-tree.
14:10:03 Join B4gder [0] (
14:13:44markunI'll just do the Gigabeat for now then
14:15:52JdGordondid anyone have any problems with me moving the filetypes array out of tree.c?
14:16:36linuxstbTo where?
14:17:13JdGordonit fits in there more than tree.c imo
14:20:16LinusNi have no objections
14:20:43linuxstbWouldn't it also make sense to rename the TREE_ATTR_* #defines to something else, and move them to filetypes.h ?
14:21:06JdGordonpossibly... but that means fiddling with more than just 4 files :p
14:21:10linuxstbi.e. filetypes should be depending on code in tree...
14:21:37JdGordonI can move the #defines across also, but I dont want to rename them
14:22:41amiconnhaha, /me wonders who defined TREE_ATTR_RWPS and up
14:23:08markunout for luch, my commit shouldn't break anything..
14:23:34linuxstbJdGordon: It just seems the intention of your commit is to abstract the filetype handling away from the tree code, but you've only gone half way...
14:23:42JdGordonamiconn: ?
14:24:14JdGordonlinuxstb: ok, your right... FILE_ATTR_?
14:24:36linuxstbamiconn: svn annote blames someone called dave....
14:24:42amiconn0x0900 -> 0x1000: 6 IDs wasted...
14:25:31JdGordonsomeone forgot which base they were in :D
14:25:32B4gdersvn blame is a great alias for annotate ;-)
14:25:45 Join pearldiver [0] (
14:25:46linuxstbJdGordon: Maybe just FILETYPE_MPA etc ?
14:26:44JdGordonno, I think it should keep the ATTR, to remind you its |'d with a real value
14:27:09JdGordonbah i dunno
14:30:18JdGordondoes anyone know if the values for FILE_ATTR_* is checked without the define? may as well fix the missing values unless they are checked wrongly...
14:30:35linuxstb"svn blame" lies - all I did was change some whitespace in the incorrect line, this is the original commit that added it -
14:30:36amiconnThe TREE_ATTRs are used in the tree. So why rename/move them?
14:30:48*linuxstb can go back to work now
14:30:54amiconnAnd then there's the thumbnail attribute which isn't really related to a type
14:31:16 Quit brun0_ (Read error: 104 (Connection reset by peer))
14:32:07JdGordonamiconn: they are used by onplay.c, filetree.c and tree.c,
14:33:03 Quit rift_ (Read error: 54 (Connection reset by peer))
14:35:41JdGordontree.c uses them 4 times.. filetree.c uses them 4x that
14:36:08JdGordontree.[ch] is one of my current pet hates :p
14:37:06JdGordon43 times in filetree.c
14:37:54 Quit pill (Nick collision from services.)
14:38:54 Join _pill [0] (
14:40:45 Join BigBambi [0] (
14:41:53 Join Soap2 [0] (
14:46:27JdGordonnew patch
14:46:28amiconnpondlife: Hmm, can't make it crash here...
14:47:40 Join rift [0] (
14:48:13linuxstbJdGordon: Seems you have an unrelated change in gui/icon.c
14:48:46JdGordonta, missed 1 file cleaning it then
14:54:01JdGordonlinuxstb: look ok?
14:54:14linuxstbJdGordon: Patch seems logical to me. Maybe FILETYPE_MPA should be renamed to something generic (e.g. FILETYPE_AUDIO) if it is in fact used for all supported audiotypes. But that can be done some other time...
14:54:49JdGordonmay as well do it now... easier now anyway seen its only the patch which needs fixing
14:57:04JdGordonanything else?
14:57:44 Join Arathis [0] (
14:58:05linuxstbJdGordon: No more comments from me...
15:04:10*JdGordon hates to think how slow rockboy would run if you have music decodnig in the background :p
15:04:27 Quit BigBambi (Read error: 110 (Connection timed out))
15:10:48JdGordonyay, no nasty surprises from the commit :)
15:11:05 Join Hammer89 [0] (
15:11:24JdGordonhow does Slasheri get away with all the yellow?
15:11:33B4gderhe keeps away from here ;-)
15:11:56 Quit Jon-Kha (Remote closed the connection)
15:12:45***Saving seen data "./dancer.seen"
15:14:59*linuxstb is quite impressed with tak, a new lossless format
15:16:04linuxstbSeems to compress to similar levels as the high Monkey's Audio settings, but decodes in a reasonable time...
15:16:32hcssimilar to FLAC, from what I'm reading, cool
15:16:53hcs(for decode speed)
15:16:53 Join Jon-Kha [0] (
15:17:12linuxstbMy tests show it's about half the decoding speed of flac.
15:17:23linuxstbI mean double the time - half the speed...
15:17:44LloreanIs it open source?
15:17:58linuxstbNot yet...
15:18:22linuxstbIIUC, the author has written it in Delphi and won't release that source, but is planning an open source implementation in C.
15:22:23 Quit Jon-Kha (Remote closed the connection)
15:23:15B4gder"TAK will be open-source, as soon as the code is ported to C or C++ and documented."
15:23:27B4gderlet's hope he does the right choice there then...
15:24:04LloreanMaybe a polite email suggesting C would be beneficial? ;)
15:24:07Zagormakes you wonder what "gems" are hidden in the delphi code
15:24:37B4gderhere's a fun comparison chart:
15:25:27pondlifeamiconn: Hmm, I can make it crash without the patch, so false alarm.
15:26:02hcsExtra, Extra Extra, Extra Max... you'd think the natural measure a codec developer would use would be numeric
15:26:29LloreanTak does look really good from that chart.
15:26:37B4gderhcs: remember this is a delphi guy ;-)
15:27:36 Quit himitsu (Read error: 60 (Operation timed out))
15:28:10ZagorLlorean: does it? 2% better compression than flac at same speed
15:28:19pondlifeamiconn: It doesn't crash every time, but repeatedly going into JPG viewer with music playing will kill the sim.
15:28:28LloreanZagor: For Lossless, that's pretty good.
15:29:00 Join Jon-Kha [0] (
15:29:08Zagorwell sure, but not nearly good enough to stray from the "standard" imho
15:29:23LloreanI mean OptimFrog only gets 1% better, at 1/22nd of the speed.
15:29:38LloreanYeah, FLAC is good enough for me, I'll agree.
15:29:51linuxstbI agree, it's going to take more than a couple of % to make me move away from flac.
15:32:27 Quit aliask ("ChatZilla [Firefox]")
15:36:45 Join miepchen^schlaf [0] (
15:37:11 Quit Jon-Kha (Remote closed the connection)
15:39:40 Join Jon-Kha [0] (
15:41:53JdGordon have a listen to that with headphones on...
15:42:04JdGordonvirtual haircut... very well done
15:43:46 Join docwhat [0] (
15:44:13 Quit Entasis (Read error: 104 (Connection reset by peer))
15:44:16docwhatHow do I change the color of the icons and the 'cursor' in the theme? I can't see black on dark....
15:44:44Lloreandocwhat: You'll have to use a custom icon set.
15:44:55docwhatOh! there are icon sets?
15:45:07LinusNJdGordon: rapidshare my a**, endless clicking to get the damn file :-/
15:45:14docwhatWhere can I get them?
15:45:23Lloreandocwhat: The IconSets wiki page, or create your own
15:45:54docwhatAh,h hah. Under extras....thanks Llorean. :-)">
15:47:38 Quit juxtap (Read error: 110 (Connection timed out))
15:47:38amiconnImho "rapid"share doesn't deserve its name
15:49:11JdGordonit used to... before they sold out :p
15:50:04 Part LinusN
15:50:52 Join Febs [0] (n=chatzill@
15:52:54*amiconn wonders why JdGordon's commit increased binary size for arm targets - only
15:53:05preglowbecause arm is special
15:53:50JdGordonhmm... aparently i forgot to refresh the build page.. I thought there was 0 delta
15:54:08 Join kkurbjun [0] (
15:54:20JdGordonthe only thing that could have done that was the addition of config.h to tagtree.c (which I did because it should always be included shoudnt it?)
15:54:28*B4gder likes when there are 5 Haxx servers in the top of the build server stats ;-)
15:54:30 Quit pondlife ("disconnected has pondlife")
15:55:54 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
15:56:13 Join miepchen^schlaf [0] (
15:58:30 Join funky [0] (n=repulse@unaffiliated/funky)
15:58:33*JdGordon wonders why none of the file*.c's include config.h
15:59:02 Quit webguest06 ("CGI:IRC")
15:59:35 Quit Nibbier (Remote closed the connection)
16:02:59 Join Nibbier [0] (
16:04:45 Quit B4gder ("Lämnar")
16:05:03 Join desowin [0] (
16:06:24 Quit JdGordon ("Konversation terminated!")
16:07:27 Join My_Sic [0] (
16:08:13 Join cybrhuman [0] (
16:10:07 Quit Topy44 (Read error: 104 (Connection reset by peer))
16:10:25 Quit docwhat (Remote closed the connection)
16:12:42 Quit Jon-Kha ("leaving")
16:15:25 Part cybrhuman
16:15:35 Join ceaser_ [0] (
16:16:18 Join Jon-Kha [0] (
16:16:29 Join darkless [0] (
16:16:53 Quit Haarg ()
16:17:39parasite_Hi, I was wondering if it is possible in any way to display the battery voltage on the WPS
16:22:16PaulJamsure, %bv
16:22:45 Join Topy44 [0] (
16:23:12parasite_thank you very much
16:23:25*parasite_ blushes about asking a question with an answer on the wiki
16:29:04markunwell, you just have to know where to look
16:30:43Topy44a silly idea, but since there is a "simulator" it should probably be very easy to do: has anyone ever attempted to make a x86-based standalone DAP based on rockbox?
16:31:17 Quit Siltaar (Read error: 60 (Operation timed out))
16:31:35markunTopy44: not really, but there has been talk about it from time to time
16:31:36GodEaterx86 isn't an ideal embedded platform
16:31:42markunGodEater: so?
16:31:56Topy44nope, but just about everyone has some old hardware standing around anyway
16:32:05GodEatermarkun: so based on that I don't believe anyone's tried it ?
16:32:26markunah, I misread the question :)
16:32:57Topy44display output over a small 7" tft or a parallel port LCD, controlls over a hacked keyboard and/or an infrared remote, ...
16:33:00GodEaterfor DAP I think in terms of something portable too
16:33:16GodEateryou could make a kind of hifi unit from it I suppose
16:33:21Topy44thats what i meant
16:33:29GodEaterright - sorry I misunderstood
16:33:37GodEaterI thought you meant something portable
16:34:09Topy44would be great for a car unit aswell... for example using an epia board
16:34:25*GodEater wishes empeg were still going
16:34:30GodEaterI always wanted one of their head units
16:34:42GodEatercurse Rio!!!
16:34:53Topy44i might actually attempt this... i always wanted a standalone hifi mp3 player
16:35:18Topy44i already built hardware for it once (fitting a pc into the case of an old cd player)
16:35:34Topy44but i never properly set up any software for it
16:35:57Topy44does the simulator output sound?
16:36:06Topy44(i havent used the simulator yet)
16:36:36linuxstbTopy44: Lots of people have talked about using the simulator in various situations (i.e. other than debugging Rockbox), but I don't think anyone has done it.
16:36:48Topy44i really might try this
16:39:26 Quit Jon-Kha (Remote closed the connection)
16:39:57 Join Siltaar [0] (
16:40:27 Join x1jmp [0] (
16:40:28Topy44of course, the nicest way of doing it would be to run rockbox without having a full linux behind it
16:40:51Topy44fast bootup and low resource needs
16:41:08Topy44but just a stripped down custom linux distro should do for a start
16:41:50 Join Jon-Kha [0] (
16:43:02linuxstbTopy44: The sim uses SDL for keyboard input, display and sound (yes, sound works in the sim).
16:43:46markundoes SDL run on (free)dos?
16:44:07markuns/run on/exist for/
16:44:19jhMikeSamiconn: my commit msg? yeah, guess you're right :P stupid allergy medication again ... every time I need to take it I end up doing one like that. :D
16:44:29Topy44input and display output to a custom lcd could be easily handled by a custom driver i guess... didnt look at sound programming yet, but should be easy if used with a specific sound card
16:44:32GodEatermarkun: I don't believe so
16:45:02Topy44i havent done much low level code yet
16:45:12Topy44but theres always a first time :)
16:47:34Topy44a bunch of scripts could handle mounting external media like cdrom and usb sticks
16:47:56Topy44i think the changes to rockbox itself would be minimal
16:48:20amiconnjhMikeS: Your commit doesn't change the actual code at all, but the commit msg suggests it does
16:48:59linuxstbCan svn log messages be changed? I think I read somewhere that they can't.
16:49:01jhMikeSamiconn: Yeah, I know. I woke up, checked the logs and looked again and realized what I'd done. oy
16:49:23jhMikeSIf Bagder installs the hook for it, yes
16:49:40Topy44btw, where are you guys from? is there any regional concentration of rockbox developers or are you spread all over the world? :)
16:50:38markunTopy44: a lot are from europe
16:50:44markunlet me try to find a map
16:50:52Topy44i'm from germany">
16:52:01*GodEater doesn't show up on that map so points vaguely at London, UK
16:52:42markunand there are quite some devs on that map I've never heard of..
16:55:01Topy44hm, i just tried setting up the simulator... everything seems to have worked, no errors, but the /sim/archos directory is empty
16:56:35jhMikeSIf we're going to do subs in mpegplayer, then I think we need to parse out the program stream time. What's it? PTC or something? hrm
16:56:39markunTopy44: "make install"
16:56:51Topy44markun, did that of course
16:57:11linuxstbTopy44: There should be a ".rockbox" directory - which may not be shown depending on how you're looking...
16:57:37 Quit Zagor ("Client exiting")
16:57:59Topy44me idiot. yes, of course
16:58:00linuxstbjhMikeS: PCR?
16:58:36 Join BigBambi [0] (
16:58:46jhMikeSlinuxstb: I guess that's it. They're suppose to be there for that purpose.
16:58:47linuxstbAh, no PCR is in transport streams, SCR is in program streams
16:58:55linuxstbSystem Clock Reference
16:59:00 Join toffe82 [0] (
16:59:19 Quit BigBambi (Read error: 104 (Connection reset by peer))
16:59:20linuxstb"A timestamp in the Program Stream from which decoder timing is derived"
16:59:33LloreanAre you planning on doing subs?
17:00:03LloreanIt seems like that would be an encoding step, for portable use.
17:00:11jhMikeSThat would also keep audio synced properly if audio frames are dropped instead of rushing the video ahead to keep up.
17:00:55 Join midgey [0] (
17:01:16linuxstbWhy would you drop audio frames?
17:01:31jhMikeSif you have errors
17:03:35Topy44well, personally, i would like having softsubs
17:03:37jhMikeSright now the whole clock will just jump ahead but actually silence should be played where an audio frame is bad
17:04:02Topy44since subtitles use up a lot of video bandwidth because of their sharp contours and high contrast
17:04:10 Quit sando (Read error: 54 (Connection reset by peer))
17:04:41Topy44they would be more compact and better readable as softsubs
17:05:23Topy44anime episodes are predestined for portable viewing, as they are short, dont suffer to much by the lower framerate and dont need very high bitrates
17:05:37 Join vcardenas [0] (i=c8767629@gateway/web/cgi-irc/
17:05:47Topy44with softsubs, that would be a real feature
17:06:26Topy44(i am regularly using the X5's original firmware to watch cartoons on the go already)
17:06:29jhMikeSright now, I don't really know the technical difference. One is another MPEG video stream and softsubs are what exactly?
17:06:38Topy44text with timestamps
17:06:39markunjhMikeS: text
17:07:05LloreanSoftsubs look better, but need some sort of overlay.
17:07:05Topy44(or overlay images in the case of dvds, but thats impractical for portable use)
17:07:24*dionoea has subs working in mpeg player
17:07:30dionoeasee yesterday's log for more info
17:07:38dionoeajhMikeS: were you working on subs support ?
17:07:51markundionoea: I was trying to remember your nick to mention it :)
17:07:56jhMikeSIf I get the framedropping tuned up I can hit about 16-17fps regulated on x5 with elephants dream
17:08:22dionoea(using srt txt subs)
17:08:27*Llorean thinks 15fps is usually "good enough" for video, for him.
17:08:27jhMikeSdionoea: no, but saw your questions in the logs
17:08:32Topy44screen size could really be an issue with subs though, longer sentences might just clutter up the whole screen with text...)
17:08:46dionoeaI'll post a patch when i get back from work
17:08:56Topy44still, it would be a nice feature imho
17:09:00markunTopy44: and the X5 in particular of course
17:09:12 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
17:09:20LloreanTopy44: With some cleverness one could possible look at the next sub's timestamp, and scroll the sub across the screen ending 1 second before the next one, or something similar.
17:09:34amiconnjhMikeS: Is that with miark's code?
17:09:34*jhMikeS needs to get dropping really optimized but it doesn't take well to simple math equations
17:10:20Topy44what was the original framerate of the video file? (as in, how much did you have to drop at decoding time?)
17:10:28jhMikeSjust with more optimal choice of dropping frames in the decoder
17:10:57jhMikeS24 fps
17:11:19jhMikeSmostly b frame drops which really benefit it
17:11:23Topy44simulator is working for me now btw, but you should change the tutorial, since the file is not called rockboxui.exe but just rockboxui
17:11:31linuxstb_jhMikeS: Why is frame-dropping so important to you? Wouldn't it be better to simply encode files at a framerate the target device can actually decode?
17:11:48LloreanTopy44: It's rockboxui.exe in windows.
17:11:55LloreanTopy44: Which tutorial are you referring to?
17:12:20Topy44Llorean, aah, didnt think about that :)
17:12:28Lloreanlinuxstb_: There's a range between 15 and 24 that doesn't have a good framerate choice though, right?
17:12:30jhMikeSbecause it should be tuned to the hilt really
17:12:31Topy44 <== setting up the simulator
17:12:36LloreanSay a target can do ~21fps
17:12:49***Saving seen data "./dancer.seen"
17:12:56LloreanTopy44: Ah, yes, that's a user created guide, that some of us aren't entirely fond of. ;)
17:13:05Topy44ok :)
17:13:10LloreanOr at least I'm not. It seems a little much to be called "simple"
17:13:15jhMikeSIt might even be possible to acheive higher framerates by dropping selectively than what it can do with no regulation
17:13:32Topy44personally i find the tutorial quite good
17:13:51jhMikeSand if it must drop, it should look as even as possible
17:14:06markunTopy44: if you create a wiki account and ask for write permission you can fix the mistakes you find
17:14:15Topy44ok, i will
17:15:00LloreanTopy44: It's got a lot more text than is necessary, or did last time I checked, and seems to be a bit patched together. But that's just my personal opinion
17:15:32Topy44but it contains everything needed in a understandable way
17:15:53 Join Juice- [0] (n=Juice@
17:15:54Topy44i will clean it up a little
17:16:28dionoeajhMikeS: btw, was there any need for other on screen display (that is, not subs)
17:16:40dionoealike time / volume info or stuff like that
17:16:43Topy44so, i'm registred.. wikiname ThorinHopkins
17:16:46dionoeaoverlayed on the video
17:16:49Topy44could i have write permission
17:16:53LloreanTopy44: There's a lot of unnecessary stuff to. For example, there's never really a good reason to use the Daily or Bleeding Edge source code. Anyone *should* use the SVN source.
17:17:08markundionoea: a seek bar I would say
17:17:13Lloreandionoea: Volume might be nice as well
17:17:43Topy44a FPS display for debugging/finding out what the specific target is capable of might be nice
17:17:49dionoeathe only issue is that i'm currently rendering in the vieo buffer, which means that black bars can't be drawn into
17:18:02LloreanTopy44: Already is one
17:18:10dionoeaTopy44: that already exists (but it's not rendered optimaly)
17:18:22 Join Topy [0] (
17:18:28Topyargh, 12h disconnect by my ISP...
17:18:35markundionoea: and your code will probably look wrong on the rotated Gigabeat screen
17:18:48LloreanTopy: there already is an FPS display
17:18:54LloreanVery debuggy looking, and everything
17:18:55linuxstb_markun: And Sansa...
17:18:56dionoeamarkun: i doubt it. It's rendering in the video buffer before it gets displayed
17:19:04dionoeanot on the lcd
17:19:08markunah, ok
17:19:13 Quit vcardenas ("CGI:IRC (Ping timeout)")
17:19:19dionoearotation is done when copying the buffer to the lcd right ?
17:19:29markunwell, you mean the yuv buffer?
17:19:30jhMikeSdionoea: OSD will commence
17:19:41amiconnThe rvf videoplayer does display volume when changed, and a seeking bar
17:19:46Topyas a quick hack, couldnt you just pad out the video buffer to screen resolution before output?
17:19:55dionoeamarkun: yep (well, in the y buffer in fact. Nobody cares about u or v for b&w subs)
17:19:58Lloreandionoea: Well, then why not just put a check for 4:3 vs 16:9(ish) ratio, and only activate your code for the 4:3, and come up with a simpler method for 16:9
17:20:11amiconnIt doesn't display an overlay, but rather cuts away the bottom 8 pixels from the video display for that
17:20:13LloreanTopy: Blitting the whole screen slows down things
17:20:36dionoeaLlorean: simpler means longer memory access times (since you'll need to draw in the lcd buffer instead of in some buffer in the RAM ... unless that assumption is wrong)
17:20:38TopyLlorean, there are other aspects around, since videos found on the net might have weird aspect ratios
17:20:47Topyof course, one could pad it during encode
17:20:52 Quit linuxstb (Read error: 113 (No route to host))
17:20:58amiconnThis should be really simple with lcd_yuv_blit too: just don't blit the whole frame
17:21:07dionoeaand i'm currently rending in the y buffer (which is 1byte per pixel) instead of 2bytes for the lcd
17:21:13LloreanTopy: Yes, so you just check if screen height - video height is greater than the needed overlay height
17:21:14dionoeawhich might make it even faster
17:21:20LloreanErr, non-overlay height
17:21:35LloreanTopy: The ratio isn't really important, just how big of a black bar you can have.
17:22:05Lloreanamiconn: I like that, but the problem still remains for letterboxed videos.
17:22:15markunLlorean: why?
17:22:25Topysubs in a video overlay should be outlined aswell, white-on-white isnt very readable :)
17:22:27 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
17:22:31LloreanDrawing the volume or subtitles in the video area when you have a letterbox you can write it in instead isn't really preferable.
17:22:44dionoeabut it's faster :)
17:22:52dionoea(in terms of CPU on the DAP)
17:23:05markunLlorean: amiconn was talking about drawing at the bottom of the screen
17:23:24dionoeaSince you'll still be rendering the same yuv rectangle as when you have no OSD
17:23:30Lloreanmarkun: Yeah, it didn't click wholly in my mind.
17:23:45Lloreanmarkun: Draw at the bottom of the screen, and for full-frame video, reduce the blit area when text needs to be there, right?
17:23:53markunjhMikeS: so, what's it going to be? :)
17:24:53Topy[17:20] <Llorean> Topy: Blitting the whole screen slows down things <== would that make so much of a difference? isnt decoding what limits the framerate most?
17:25:29markunTopy: yuv2rgb was consuming 60% of the time on the Gigabeat iirc :)
17:25:34Topyoh, ok :)
17:25:52markunI don't know how much it is now (after some optimisations)
17:26:16LloreanStill enough that we don't want to blit more than we need to. :)
17:26:17jhMikeSmarkun: I suppose it can vary in terms using alpha vs. 1bit transparency
17:26:35amiconnI mean cutting away a little from lcd_yuv_blit() at the bottom if necessary, and just output the text with the normal lcd drawing followed by an lcd_update_rect()
17:26:42 Quit Topy44 (Read error: 60 (Operation timed out))
17:26:51Lloreanamiconn: The only problem being the 3:4 screen targets.
17:27:07*dionoea still thinks that rendering the OSD in the LCD buffer is a mistake in terms of performance
17:27:16Lloreandionoea: I like amiconn's method
17:27:18jhMikeSamiconn: yes, that's probably the best approach for slow targets.
17:28:07markundionoea: we can do some benchmarks
17:28:26jhMikeSI think YUV code in a plugin lib would be better since I might needs several different ones
17:28:41dionoeaWell it'd mean that: 1/ you have to render in another yuv buffer and 2/ you have to render that on screen
17:28:53dionoeaor that you have to render pixel by pixel on screen
17:29:06dionoeacompared to rendering in the same yuv buffer and that's it
17:29:35dionoeaif this were a normal computer, i'd agree that the other solution is awesome :) But this is a slow dap :(
17:29:56markundionoea: which one are we talking about? ipod?
17:29:59*linuxstb_ gives dionoea a gigabeat...
17:30:05dionoeai have the ipod video :)
17:30:10jhMikeSno, just render the overlay into the lcd_framebuffer and combine the two on the fly into the driver framebuffer
17:30:36markunand the overlay can stay the same for several frames
17:30:47dionoeamarkun: agreed
17:30:56jhMikeSbut that's for a gigabeat or sansa...for others just use lcd framebuffer and combine when writing to lcd
17:32:44dionoeaI'll make it possible to choose the rendering method
17:35:43 Quit heanol (Remote closed the connection)
17:35:54 Join heanol [0] (
17:38:23 Join Rincewind [0] (
17:40:07 Quit Arathis (Remote closed the connection)
17:41:38 Join PaulJam_ [0] (
17:51:07 Join bluey- [0] (
17:53:10 Join davina [0] (
17:55:03 Quit petur ("urgh")
17:55:26courtcSpeaking of lcd... The ipl kernel and ipodloader2 have photo type 0 lcd windowing incorrect...
17:55:39courtcDunno if you guys used that code...
17:56:38markunlinuxstb_: do you know that?
17:57:30 Quit PaulJam (Read error: 113 (No route to host))
18:00:10courtcIt's not a fatal error. In fact, it'll never be noticed if you use a full draw window every frame.
18:02:02 Quit ceaser_ (Remote closed the connection)
18:06:28amiconndionoea: Why would you render pixel by pixel?
18:07:55dionoeaIt's the 3rd solution only :) (because it would require rendering the white pixels only, and not the black ones ... and there are more black pixels than white pixels)
18:08:18dionoea(I currenty render text in white in the yuv buffer)
18:09:42dionoeaamiconn: and it also wouldn't require rendering in another yuv buffer (since you could write directly to the lcd pix by pix). But that still sounds suboptimial compared to rendering directly in the video yuv buffer
18:10:00dionoeaespecially since rendering white in yuv is simple
18:10:09 Quit lids (Remote closed the connection)
18:10:39dionoea(you only have two colors which can be rendered simply in yuv: white and black)
18:11:06dionoeaif you try rendering colors you have to handle the 4:2:0 y to u/v ratio ... and that's kind of complicated (well, more complicated)
18:11:28dionoeaso you'd be better off rendering in rgb if you're rendering out of the video frame in fact
18:11:38dionoea(and yuv if you're in the video frame of course)
18:12:59dionoeaSo i'll provide an option to either: 1/ render in y(uv) on the video buffer or 2/ render in rgb on the lcd buffer
18:13:12 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
18:13:20 Quit PaulJam_ (Read error: 113 (No route to host))
18:13:29 Join miepchen^schlaf [0] (
18:13:37dionoea(rendering in rgb, for subs, is simple since it's only a call to putsxy or something like that)
18:14:25 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
18:16:18 Join linuxstb__ [0] (
18:16:52 Quit linuxstb_ (Read error: 113 (No route to host))
18:22:53*Llorean almost wants to declare anyone using Loader2.anything as running an "unsupported build" in general.
18:22:58 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:23:02LloreanI didn't even make the connection that it can cause codec errors.
18:23:34LloreanBecause it doesn't look in .rockbox first (or at all without customization)
18:24:53 Join lids [0] (i=lds@gateway/tor/x-bad012307b4e57e8)
18:25:46courtcLlorean: actually....
18:25:52Lloreancourtc: Actually what?
18:25:59amiconndionoea: I can't see how rendering to yuv should be more efficient than cutting the yuv a little during text display and using the normal lcd rendering for the text
18:26:27Lloreancourtc: Is there a newer version of loader2 that gives preference to /.rockbox/rockbox.ipod and only loads /rockbox.ipod if it's not found, as our bootloader does?
18:26:37LloreanWithout a custom config file, that is.
18:26:40amiconnlcd_yuv_blit() does the conversion from yuv to rgb. Rendering to the exisiting lcd framebuffer renders directly to native format, saving the conversion
18:26:55amiconnAnd the amount of data actually transferred to the lcd is the same
18:27:02courtcIt loads .rockbox/rockbox.ipod in rev1427
18:27:11Lloreancourtc: And falls back if it's not found?
18:27:25courtcNah, I was too lazy to implement that.
18:27:57amiconn(comparing full-screen yuv vs. reduced yuv + small rectangle from normal framebuffer)
18:27:57courtcIf it's not found, you can use a config file because that's not the default amymore.
18:28:27 Quit obo ("KVIrc 3.2.6 Anomalies")
18:28:27LloreanEither way, use of Loader2 has caused too many strange problems over the history. =/
18:28:55courtcIf you have rev1427 of loader2, you will probably have a rockbox version at least newer than that.
18:29:02LloreanProbably, yes.
18:29:18LloreanIs 1427 available as binary in an official capacity?
18:29:32courtcnope. I commit code, I don't make releases.
18:29:55LloreanSo, basically, any normal user won't have it.
18:29:57*Llorean sighs
18:30:15*Llorean wishes people would use the official install instructions, or at least try an official install if they get strange behaviour.
18:30:23courtcWhich as far as I care, isn't my problem. ;)
18:30:36Lloreancourtc: Yes, you've made it my problem instead, thans.
18:30:39Lloreanthanks even.
18:30:47courtcNOt really.
18:31:05LloreanI get the support load from your installer not loading Rockbox properly.
18:31:11LloreanErr, your loader
18:31:18courtcI don't condone newbie use of ipodlinux, and I wouldn't think that you would for rockbox either.
18:31:27LloreanBecause they think it's our job to support your loader if they're trying to load Rockbox with it.
18:32:56markunLlorean: too bad not everyone wants to get rid of the runtime-estimation as it is
18:33:00LloreanPeople can't make the connection that if the loader is causing the problem, it's not our problem. We didn't write it.
18:33:09Lloreanmarkun: I think it's as useless as % volume.
18:33:15markunme too
18:34:10markunThe say a wrong estimation is better than no estimation, I disagree with that
18:34:10LloreanSame here.
18:34:10markun"it tells me 15 hours left so I have about 1 hour left"
18:34:13courtcLlorean: I didn't build the ipod...
18:34:37Rincewindmarkun: the runtime prediction is very accurate on my h120 I would miss it badly
18:34:44markuncourtc: but you wish you had of course, then you had all the portalplayer docs :)
18:34:47Lloreancourtc: Yes, so if the flash-loader fails (as it often does when the battery is low), it's not your problem or ours. :)
18:35:01LloreanRincewind: Can you do 80% of 15 hours in your head?
18:35:26RincewindLlorean: I don't want to do it if I don't have to
18:35:34markunyou will have to :)
18:35:35courtcRight, but people still ask for support...
18:35:42Rincewindand the runtime display is more impressive
18:36:18LloreanRincewind: It's also horribly misleading for many cases.
18:36:18markunI would like runtime estimation, but just by keeping track of the time and the percentage changes
18:36:18Rincewindmarkun: what about my suggestion about letting the user set the max runtime to make it work better?
18:36:18LloreanRincewind: It's never been right on my H120, because it has an old battery.
18:36:44Lloreanmarkun: If 1 hour has passed, and % has gone down 10%, you have 9 hours remainig?
18:36:55markunLlorean: like that
18:37:07*amiconn thinks the runtime estimation is a useful feature
18:37:21Lloreanamiconn: I think it's misleading for many users.
18:37:30RincewindI don't mind if the current algorithm is simplified a bit, but please don't remove it
18:37:43amiconnIt's just misleading because it isn't calibrated properly for a number of targets
18:37:54Lloreanamiconn: It doesn't work well with old batteries, at least not for me.
18:38:01markunRincewind: and I think the other way around. Remove it if it doesn't get fixed :)
18:38:12pixelmaLlorean: I don't care about the runtime estimation much - but the percentage is percentage of volts and not runtime - as far as I could see it's not linear
18:38:16 Join adiamas [0] (n=chatzill@
18:38:23courtcadd runtim calibration...
18:38:33Lloreanpixelma: The battery % should be based on the discharge curve, I thought.
18:38:42amiconnpixelma: Percentage is percentage of charge, i.e. proportional to runtime as long as conditions stay the same
18:38:42 Join PaulJam [0] (
18:38:51markunpixelma: it should be linear
18:39:34markunbut that's not always so easy with the values we get from the ADC
18:39:34amiconnI am thinking for quite a while about making runtime estimation (and percentage display) auto-calibrating
18:39:39Lloreanamiconn: I would be happy with that.
18:39:43markunamiconn: yes, I was thinking about that as well
18:40:28Lloreanamiconn: I want it either gone or adaptive somehow or based on a time that a user sets rather than capacity (so people don't say "Why does my X say I only have 9 hours left when I should have 16?" and "Why does it say I have 16 when I only get 10")
18:40:34pixelmahmm... maybe I'm misunderstanding but from what I can see when I follow percentage display it doesn't develop "even" (can't find a better eplanation at the moment)
18:41:48amiconnIt does, or at least should
18:42:36amiconnOn targets with NiMH batteries it won't always do, because voltage isn't a reliable indicator for the remaining charge
18:43:24amiconnThe only way to do it prroperly for NiMH would be tracking the actual charge, which would require a means to measure actual discharge current, but the hardware doesn't allow that
18:43:32Slasheripixelma: the percentage is calibrated linear amount of charge left on the battery
18:43:49Slasheriat least it should be, not on ipods necessary and so on
18:43:50amiconnFor LiIon, voltage can be mapped reliably to the remaining charge
18:44:05spiorfsorry, but i throw my idea in the middle. couldn't rockbox read the battery_bench output and calibrate voltage/percentage/time left based on those infos?
18:44:06amiconn(and LiPo of course)
18:44:36amiconnspiorf: It would need to know the conditions of the battery_bench run
18:45:02amiconnRuntime estimation isn't just remaining percentage multiplied by a fixed value
18:45:37amiconnIt tries to take the (varying) discharge current into account
18:45:39spiorfamiconn, it's always an "estimation"
18:45:46pixelmaso I see... it's also only based on my personal impression, I would have to note it down... (and yes I'm using NiMH most of the time)
18:46:11Slasheri(but it almost it, it's not quite smart to take account the average discharge current)
18:46:44spiorfimho it would be the best general method. if my player can run for 12 hours it will never tell me 20h left.
18:47:11spiorfobviously i should run baterybench at least one time
18:47:38Rincewindby the way: what is the graph in the battery entry in the debug menu telling me?
18:48:08spiorfvoltage i think
18:48:47Rincewindspiorf: I think it is more than that, because it is a block diagram, not a single value
18:49:12spiorfRincewind, the graph ov the voltage vs time
18:51:16 Join Everybody [0] (
18:51:27courtchey everybody.
18:51:33courtchad to. sorry.
18:51:44markunbecause it's so quiet?
18:51:57courtc... no.
18:52:43 Join Guile [0] (n=Guile@
18:53:04toffe82the end of the internet radio :
18:55:34spiorfanyway, it would be very simple, at least for non nimh. it looks in the batterybench log the voltage range in which the current voltage fits, then estimate time left and percentage based on how long the player lasted when it hat that particular voltage.
18:56:22Lloreanspiorf: That assumes that you're doing the same thing as when the battery bench happens.
18:56:28LloreanEven music format can change things
18:56:53spiorfif i use mp3s, batterybench will run with mp3s
18:57:04spiorfsame thing for other formats or even mixed formats
18:57:18RincewindLlorean: we can never predict a good estimate for every kind of music
18:57:23spiorfyeah, video and games will use more battery, but everyone knows that
18:57:25LearOuch. Got an icon-drawing segfault in the sim...
18:57:30 Quit Everybody ("( :: NoNameScript 4.02 :: )")
18:57:56LloreanRincewind: Yes, which is why I'm not a fan of giving people an estimate of "Time" remaining, rather than just using a percentage, which is valid under any conditions.
18:58:08courtcdetermine the voltage loss in a particular span of time.
18:58:13LloreanIf you show people a time, for some strange reason they expect it to be right.
18:58:15spiorfit would be something like "if you stop playing and watching videos right now, you can listen at 3hours and 20minutes of music
18:58:15Rincewinda percentage is not valid under every condition
18:58:30LloreanRincewind: Assuming the percentage is calibrated, under what condition isn't it valid.
18:58:52Rincewindif the percentage is calibrated, then runtime is exact, too
18:58:56LloreanNo, it's not.
18:59:02spiorfLlorean, the problem is that there is no absolute method to know the battery left.
18:59:27spiorfmy player can't know if in the next hour i will play a video or a mp3
18:59:29LloreanRincewind: If your battery is old, if you're using FLAC, or under a variety of other conditions, it's not very correct at all
18:59:47RincewindIf people complain that the estimate is not exact to the hour and minute, then we can only tell them not to be utopic
19:00:19RincewindLlorean: in that case I can run the battery bench in this conditions and get an exact result the next time
19:00:23courtcthe voltage loss over a span of time... used in collaberation with a calibrated voltage estimate / time left.
19:00:33Rincewind(if we do what spiorf suggests)
19:00:46LloreanRincewind: And it's still only valid if they don't change anything about their usage habits.
19:00:48 Join kahn [0] (
19:00:57LloreanRincewind: For example, I listen to NSF, Audiobooks, or encoded music in FLAC and MP3
19:01:02Lloreaneach of those has different lifetimes.
19:01:17courtcAre you seriously still going over this?
19:01:28kahnhi guys, quick question is it possible to patch using the.patch files on the tracker?
19:01:31spiorfLlorean, another solution would be to log the battery every time
19:01:45spiorfso that the next time it have updated values.
19:01:59Rincewinda user who uses a variety of formats should know that runtimes change with format and an estimate is an estimate, after all
19:02:00spiorfif i change listening habits, it will know very soon
19:02:06Lloreanspiorf: That then gives you an average. Underestimating for NSF, and overestimating for FLAC
19:02:07 Quit matsl (Remote closed the connection)
19:02:21LloreanRincewind: Then why does an estimate even matter?
19:02:31LloreanWhy not just use a percentage, if you're saying "it's okay for it to be meaningless"
19:02:49spiorfLlorean, an estimation based on the last 5 logs?
19:03:02Lloreanspiorf: Then it's an average of them.
19:03:03*courtc makes a battery time calculation for ipodlinux.
19:03:07spiorfthis thing is becoming more complicated than it should be
19:03:07LloreanIt's still not that useful
19:03:21RincewindLlorean: because it gives you a figure how much is left under your usual playing habits. If I listen to flac, and my runtime is calibrated for mp3s, then the estimate will go down slower (because the battery lasts longer)
19:03:28spiorfit's still better than an estimation based on nothing
19:03:28LloreanAll I'm saying is that runtime is inherently inaccurate, and it doesn't really serve a practical purpose
19:03:44LloreanRincewind: The battery lasts shorter with FLAC
19:03:54courtcJust for fun. Then I'll make one for ccOS; then I'll shred them both and laugh evilly.
19:03:55 Join petur [0] (n=petur@rockbox/developer/petur)
19:04:01LloreanRincewind: But if it goes down faster, that makes the VALUE meaningless, only the rate is important: Again, % is better for judging rate
19:04:04Rincewindthen turn my case around, it is still valid
19:04:33Rincewindeven an inaccurate value is meaninful for humans
19:04:46 Quit lids (Remote closed the connection)
19:05:13RincewindIf I tell you "I meet you in about an hour" that's ok for you, isn't it?
19:05:31courtc... arguing for the sake of being right.
19:05:33 Quit kahn ("Statistics show only 3% of imigrants are australians")
19:05:41Lloreancourtc: Arguing because I want the feature removed or fixed.
19:05:54LloreanRincewind: And if you say "about an hour" and show up in two, I'll be irritated.
19:06:13courtcI just mentioned a way to 'fix' it.
19:06:15LloreanMy H120 player says, at 75% capacity, 12hrs 30min left. Its battery only ever lasts 10
19:06:24LloreanThat's a significant percentage wrong
19:07:06RincewindLlorean: again, this can be fixed with a variable max runtime. No need to remove the whole thing
19:07:25LloreanI said I was okay with letting the user input their expected max runtime.
19:07:31LloreanAs I KEEP SAYING, I want it fixed or gone
19:07:34LloreanBut I don't like the current one at all
19:07:40LloreanAs it's not useless in a practical sense
19:07:45LloreanVery much like % volume was not useful in a practical sense
19:07:50courtcI dunno, in OSS ususally the person who want something fixed is the one who needs to do the work.
19:08:14Lloreancourtc: And I can do the work, if it comes to that.
19:08:21LloreanI can even commit it, and have people yell at me, if I wanted to be spiteful
19:08:22Rincewindin fact, at the moment I have my editor open and grep the source to find out where the calculation is done
19:08:29LloreanBut for me, "doing the work" would be removing it.
19:08:35LloreanIf someone has a better way, they should do it.
19:08:35pixelmaLlorean: it's not useless? ;P
19:08:47Lloreanpixelma: What's not useless?
19:08:48courtcThat's not a solution though.
19:08:58LloreanOh, it's not USEFUL
19:08:59LloreanI can't type
19:09:07Lloreancourtc: What's not a solution?
19:09:23Lloreancourtc: Removing an impractical feature to recover binary size isn't a solution?
19:09:41courtcdeleteing the somewhat buggy software.
19:09:54LloreanIt's not somewhat buggy.
19:10:06courtcDamn, firefox doesn't open X page correctly, delete.
19:10:10LloreanIt does what it's supposed to do, just what it's supposed to do is not useful in a practical sense.
19:10:11LloreanIt's fluff
19:10:11RincewindLlorean: impractial is a subjective view in this case
19:10:39LloreanRincewind: It only gives useful information under a very limited set of circumstances.
19:10:50LloreanRincewind: While % gives useful information for any usage habit.
19:11:30Rincewindthen we should improve it, not remove it. If we remove every feature that works only in limited circumstances then we have not much left in rockbox
19:12:23 Join lids [0] (i=lds@gateway/tor/x-bb5c9491aab4b0af)
19:12:29LloreanRincewind: Are you saying we should never, ever reconsider the worth of a feature?
19:12:50***Saving seen data "./dancer.seen"
19:13:47RincewindI didn't want to generalize that much, besides, this discussion is pointless, both of us said our opinions and nothings gonna change them...
19:14:07LloreanMy opinion can be changed
19:14:17LloreanYou just haven't said anything that I hadn't already considered before stating mine.
19:14:29LloreanIf yours can't be changed, that's rather narrow-minded.
19:15:21Rincewindmy opinion can be changed, it is just that in this case I like the runtime thing very much and use it as the only battery meter on my player
19:15:55LearArg, smells like memory corruption...
19:15:57LloreanThen feel free to code a solution that makes it work properly.
19:16:12LloreanBut if nobody's going to bother to fix it, I will continue to attempt to get it removed.
19:16:17LloreanBecause right now it does more harm than good.
19:16:42RincewindLlorean: I try to implement my little solution I posted to the mailing list
19:19:47LloreanWe'll see how well that can work then. As I said, I wouldn't object to a fix.
19:20:33Rincewindat first I have to find it in the source...
19:21:49 Join chelli [0] (
19:23:31 Quit jhulst (Remote closed the connection)
19:25:18amiconnRincewind: The battery graph is telling the voltage over the last 2 hours, measured once per minute
19:25:58Rincewindok, that explains why there is sometimes something to see and sometimes not (after a fresh start)
19:26:24amiconnLlorean: Btw, runtime will vary less with format than you might think. (video being an exception, also because of my next point)
19:26:34hcsyay, I just won an ipod nano
19:26:38hcsboo, it is 2nd gen
19:26:44amiconnA very big sucker especially on colour targets is the lcd backlight
19:27:00 Quit lids (Remote closed the connection)
19:27:01Lloreanamiconn: On my H120 the runtime has never been accurate, and it's gotten worse as my battery has gotten bad.
19:27:17amiconnWell, it can't be correct if the battery is bad.
19:27:19Lloreanamiconn: I had attributed this mostly to my use of FLAC, but it could just be my battery's medium-condition to start with.
19:27:24amiconnThat's not the algorithm's fault
19:27:38LloreanWell the battery was getting about the right time in Rockbox initially, until I put my FLACs on.
19:27:52*Llorean shrugs
19:27:56Rincewindamiconn: the algorithm is in powermgmt.c right?
19:27:58amiconnThat's why I am pondering auto-calibration, but that's far from trivial
19:28:28amiconn(1) The user might plug or unplug the charger mid-session
19:28:29LloreanI just don't like an estimate that tells people a number of hours remaining, when it can only be right under "ideal" conditions.
19:28:59amiconn(2) In usb mode we can neither predict runcurrent, nor the charge current we get from usb
19:29:08 Join Arathis [0] (
19:29:14LloreanIf I look at it and see 12 hours remaining, that information might or might not be useful. If I looked at it 1 hour ago and saw 14 remaining, then that 12 is more useful. =/
19:29:31 Quit barrywardell (Remote closed the connection)
19:29:53amiconn(3) The targets where the batteries *can* be changed (archos player, recorder v1) or even *have to be* changed regularly (Ondio, iFP 7xx) create another problem:
19:30:05 Join oKtosiTe [0] (
19:30:44amiconnWe can't simply take averages because the new batteries can be a different brand/capacity/whatever, and to further complicate the matter, can be either alkalines or NiMH cells
19:31:22RincewindI think the minutes from the estimate can be dropped, because when it says "16h 15min" this implies an accuracy that is simply not possible
19:32:14amiconnBut otoh I do want an adaptive algorithm *especially* for recorder v1, i.e. software charging
19:32:41markunamiconn: maybe have it just for the fixed battery targets then?
19:32:44amiconnWith newer cells the algorithm tends to stop charging prematurely, giving sub-optimal runtime
19:33:16amiconn...and it even looks like that doesn't do any good for the cells
19:34:40 Join ross [0] (
19:35:58 Quit PaulJam (".")
19:36:33rossHey all. Can anyone update me on the status of support for the 5.5 generation 80 gig Video iPod? I came here over a month ago and someone told me that it was a short matter of time until there would be support because the issue had been figured out. If there are issues, is there anything I can do to help?
19:37:35 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
19:38:23 Join SirFunk [0] (
19:39:30LearGah, yet another case of enablind logf causing problems for devices that supports remotes... :/
19:41:00 Quit ross ("Leaving")
19:41:57amiconnmarkun: I think tracking the current runtime together with current approximation (including charger connection etc) and runtime-derived percentage would work best
19:42:31amiconnOn the LiIon targets looking at the voltage could be used as an additional hint, on the NiMH targets this would simply be left out
19:47:35amiconnRuntime tracking will require to see a full discharge at least from time to time - and because battery usage instructions for laptops usually say that a full discharge is required from time to time I think that's what's used there too
19:54:59 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
19:55:08 Quit amiconn (Nick collision from services.)
19:55:09 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
19:56:27 Join juxtap [0] (
19:58:13 Join elinenbe_ [0] (n=elinenbe@
20:07:17 Quit Juice- ("Leaving")
20:07:25 Join PaulJam [0] (
20:08:00 Join jhulst [0] (n=jhulst@
20:12:31 Join petit-lama-tout- [0] (
20:18:57 Nick linuxstb__ is now known as linuxstb (
20:19:39linuxstbcourtc: What was the ipod lcd bug you found? Have you committed any fixes?
20:20:06 Join mirak [0] (
20:22:52 Join bawb2 [0] (
20:25:08lostlogichas anyone ever attempted a rio carbon port?
20:25:20 Join entheh [0] (
20:25:25 Join moos [0] (
20:26:08 Join Alonea [0] (
20:26:12linuxstblostlogic: I can't remember anyone mentioning it, at least not recently.
20:27:02lostlogicmy boss has one and was wondering −− seems ot have a general purpose cpu and a microdrive.
20:29:35 Quit rift (Read error: 110 (Connection timed out))
20:30:21elinenbe_lostlogic: rio carbon = crud!
20:30:34elinenbe_that company self-destructed...
20:34:56 Join raphi [0] (
20:36:01 Join kaaloo [0] (
20:36:20 Part kaaloo
20:36:44LearIn case anyone has missed: Rockbox on About to download it myself...
20:38:17 Quit Alonea ("ChatZilla [Firefox]")
20:38:42LearAbout computer/tech stuff.
20:40:50 Join Alonea [0] (
20:41:23Aloneaok, I updated my rockbox (its been a week or two) and now all of my videos say "Incompatible Version"
20:42:21LloreanAlonea: That means your plugin version is different from your binary
20:42:49LloreanDid you install by simply extracting to the device?
20:42:52AloneaLlorean: ok. I get plugin version, but binary?
20:43:15LloreanAlonea: Rockbox.whatever (ipod, mi4, iriver...)
20:43:22LloreanWhat player do you have?
20:43:35AloneaLlorean: I compiled, made zip, extracted zip and copied it over to my drive. (have gigabeat) and overwrited what was there
20:44:16LloreanIs your bootoader up to date?
20:44:28AloneaLlorean: should be. Unless there is a new one again
20:44:47LloreanAnd is there a rockbox.gigabeat in your device root, in the /.rockbox folder, or both?
20:44:47AloneaHas it been updated last couple weeks?
20:45:06Alonearockbox.gigabeat is just in /.rockbox
20:45:45LloreanThis suggests somehow mpegplayer.rock didn't get overwritten.
20:45:49LloreanDo any other plugins work?
20:45:58Alonealet me check
20:47:13 Quit miepchen^schlaf (Read error: 113 (No route to host))
20:47:13AloneaNevermind...I fixed it
20:47:27LloreanForgot to shut down and reboot after copying over the new version?
20:47:39AloneaIt MIGHT help if I reset my player. I am sick today so brain is slow
20:47:40 Join miepchen^schlaf [0] (
20:47:53LloreanNo worries I've done it a few times too. :)
20:47:59LloreanIt'd help if RoLo worked
20:48:11*Llorean goes away.
20:48:26Aloneaoh! one of the screens is messed up a bit now
20:49:01Aloneawhen you hold down the menu button that one where it controls shuffle, repeat and show Files? The Shuffle on the left side is higher than the repeat on the opposite side.
20:49:24LloreanIt's always been this way
20:49:35LloreanOr more specifically, it's been this way since long before the Gigabeat port existed
20:49:42AloneaLlorean: its always been perfectly centered though
20:50:13LloreanRepeat has always been lower than Shuffle
20:50:17 Quit bluebrother ("leaving")
20:50:27LloreanIt's often not noticed with smaller fonts.
20:50:38LloreanOr just not noticed.
20:50:45AloneaLlorean: ah, ok. maybe thats it..its quite noticable at the moment
20:51:00LloreanTrust me, it's been this way for quite some long time.
20:51:28LloreanIf I recall it was the change that allowed the quickmenu to be in other languages, but I'm not 100% certain that was the change that did it.
20:51:44LloreanIf I recall some of the lines were too long when they were even.
20:52:24AloneaLlorean: ah. ok. Well, it just seems at more of an extreme than it used to.
20:52:59Aloneawell, I should goto class...bah.
20:53:05 Quit Alonea ("ChatZilla [Firefox]")
20:58:17 Quit x1jmp (Read error: 113 (No route to host))
20:59:37 Join Llorea1 [0] (
21:01:07 Part raphi
21:07:19 Quit linuxstb (Read error: 110 (Connection timed out))
21:07:45 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
21:09:58 Quit RaRe (Read error: 113 (No route to host))
21:12:45dionoeaanybody here know bidi_l2v well ?
21:12:53***Saving seen data "./dancer.seen"
21:16:23 Quit Llorean (Read error: 110 (Connection timed out))
21:16:39 Part pixelma
21:18:55 Join lee-qid [0] (
21:19:48 Nick dionoea is now known as bork_bork_bork (
21:20:18 Nick bork_bork_bork is now known as dionoea (
21:23:42Topydoes anyone have the pinout of the X5 subpack connector?
21:28:56 Join matsl [0] (
21:30:34 Quit desowin ("use linux")
21:35:23 Join pit_ [0] (
21:36:08pit_help! my iriver 320 doesn't boot into rockbox, it just shows a white screen
21:36:24pit_i can't reset and i can't switch it off
21:36:36pit_forgot to release the usb-cable
21:36:51pit_petur: kind of bright grey
21:36:55peturah battery flat?
21:37:17pit_petur: maybe. i tried to recharge it but the behaviour doesn't change
21:37:37peturwith th AC adapter I hope
21:37:57pit_yes. whith the original adapter.
21:38:21peturhow long did you keep it plugged in?
21:38:34pit_over night and again a whole day long
21:39:02peturthat's long
21:39:10peturand reset doesn't do anything?
21:40:33peturLinusN is your man regarding H3x0 hardware
21:40:54peturmaybe you can let the battery run flat, then try charging again
21:41:08pit_looks like he isn't here at the moment ...
21:41:16peturno he isn't
21:41:49pit_is it possible to open the hardware to charge the battery externaly?
21:41:59peturyou can open up your h320 and disconnect the battery, but it's not so easy
21:42:11peturfront and back need to be removed
21:42:15amiconnDoes that mean that even reset doesn't power off?
21:42:28amiconnNote that it won't power off if the charger is connected
21:42:31pit_i couldn't figure out how to open it - it has got 5 screws but after removing it i wasn't able go open the iriver.
21:42:37lostlogicsomeone broke tagcache disabled build on ipod 5g
21:43:13pit_amiconn: that's right. i released the charger and also the usb-cable but it wasn't still possible to reset it.
21:43:38peturpit_: here's how you open it:
21:44:10 Quit Febs (Read error: 104 (Connection reset by peer))
21:44:28pit_petur: ah great! i knew i saw it before but i couldn't find it again ...
21:44:57peturpit_: are you sure the screen is still on? maybe the battery ran flat and you hit a bug in the bootloader: try holding REC while pressing ON
21:45:53pit_petur: i tried this very often ...
21:46:57pit_i noticed that the harddisk never runs but the screen shows some lights
21:49:01 Join nls [0] (
21:49:40peturpit_: sorry, no idea... maybe try to disconnect the battery
21:51:24pit_petur: thx anyway. maybe i first try to discharge the battery by waiting one or twow days.
21:54:19amiconnDoesn't look like logf is used much these days
21:54:37*amiconn only used it 1 or 2 times back when it was relatively new
21:57:17lostlogicI used logf quite a bit when I was working on buffering bugs in playback. but that was a long time ago now.
21:58:11petur= a year ;)
21:58:19amiconnI rather place my debug output on demand (e.g. using splashes) rather than enabling logf and getting a ton of output from modules interfering with what I want to debug
21:58:28pit_thx a lot. see you!
21:58:43peturpit_: all the best
21:58:59 Quit pit_ ("Verlassend")
22:04:14LearI always enable logf in the simulator (more out of habit, but it can be handy sometimes), since I use it for debugging.
22:05:12 Join lini [0] (i=pugsley@
22:05:15amiconnImho the simulator doesn't need logf at all, since there is debugf, which is always enabled anyway
22:06:03amiconnFor the target it probably depends on what you're doing
22:06:23amiconnE.g. I didn't need any logf/debugf at all for my charcell lcd driver rework
22:06:51amiconnIf the lcd driver doesn't do what it should, it's quite obvious (and logf would probably not help either)
22:07:07 Join jonezi [0] (
22:07:28*amiconn fails to remember when he last used logf or debugf
22:07:47LearBtw, the "it" in "use it for debugging" refers to the simulator, not logf, in case that wasn't clear. :)
22:08:12joneziAnyone here know anything about the devolpment of RB for iriver h10?
22:08:56 Quit nls (Remote closed the connection)
22:11:26linuxstbjonezi: barrywardell is the main H10 person, but he isn't here. However, if you ask a question maybe someone will know.
22:11:35 Join nls [0] (
22:11:53joneziJust got a quick question about battery life and its optimization.
22:11:54 Quit matsl (Remote closed the connection)
22:12:43linuxstbjonezi: That's probably not specific to the H10, but all devices with the same CPU (ipods, Sansa)
22:13:46joneziSo as I understand it the battery has not yet been fully optimized for these devices, Coorrect?
22:13:47 Quit Thundercloud (Remote closed the connection)
22:13:48 Join hannesd_ [0] (
22:14:13 Join TrueJournals [0] (
22:14:16 Join Thundercloud [0] (
22:14:16linuxstbWell, Rockbox is draining much more power than we would expect it to.
22:14:23 Quit adiamas ("ChatZilla [Firefox]")
22:14:32LearWell, the battery isn't really something you can optimize... :p
22:14:49joneziwell, the CPU then maybe
22:14:53linuxstbThe general concensus is that we're not configuring the portalplayer chip optimally. But that's hard to fix as there is no documentation.
22:16:34dionoeahum crap, i broke my subs code ... (shouldn't have changed everything at once i guess :p)
22:17:02 Quit bluey- (Read error: 110 (Connection timed out))
22:19:40 Quit nls (Remote closed the connection)
22:20:03 Join nls [0] (
22:20:04 Join Mmmm [0] (
22:20:29 Quit nls (Remote closed the connection)
22:20:54jonezithanks for all the work you guys put in anyway, muchly appreciated, hope you solve this problem soon!!!! Cheers
22:22:53 Part jonezi
22:26:13Nico_Pdionoea: you got subs almost working ?
22:26:49 Quit Lear ("Chatzilla 0.9.77 [Firefox]")
22:27:36dionoeaNico_P: i had them working, i cleaned the code and now it's working (except that timing is a bit screwed now ... i'll find the error and fix it)
22:27:52Nico_Pdionoea: that's awesome !
22:27:53dionoeaCleaning code is always a risk :)
22:28:05Nico_Pmpegplayer is progressing very fast
22:29:43 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
22:33:00BHSPitMonkeyis there a gbs player in rb?
22:34:50peturgameboy music?
22:35:52peturstart porting one?
22:36:40peturgoogle points me to
22:37:22BHSPitMonkeyI guess that's a no?
22:37:47peturwell not that I know of...
22:40:01linuxstbnezplug is "free for non-commercial use" IIUC.
22:40:36hcsgame_music_emu is GPL, but C++, and it has a fair amount of funny band-limited synthesis
22:48:32 Join inversions [0] (
22:49:02 Quit Arathis ("Bye, bye")
22:49:59*petur kicks buildserver
22:50:56*pixelma wonders when "buildserver" is going to show up here ;)
22:51:09 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
22:51:29 Join miepchen^schlaf [0] (
22:51:42 Quit billytwowilly (Read error: 104 (Connection reset by peer))
22:51:48*petur hasn't seen forehead for a while - the irc user, that is ;)
22:51:56 Quit chelli ("Client exiting")
22:52:16 Join billytwowilly [0] (
22:52:34 Join buildserver [0] (i=52b61a05@gateway/web/cgi-irc/
22:52:50 Quit buildserver (Client Quit)
22:53:11 Join forehead [0] (i=4ca871a9@gateway/web/cgi-irc/
22:53:16foreheadyou rang?
22:53:39hcsyou wrong
22:53:53*petur slaps forehead
22:54:04foreheadin soviet russia....
22:54:11*forehead slaps petur
22:54:20dionoeaNico_P: i think that i found my bug's cause. I'll be able to commit soon
22:55:13 Quit forehead (Client Quit)
22:55:37 Join SliMM [0] (n=chatzill@
23:01:29petit-lama-tout-i have shutdown my ipod
23:01:38petit-lama-tout-and now it doesn't start
23:02:29 Join hannesd__ [0] (
23:04:32 Quit midgey ()
23:05:00 Quit Soap2 (Read error: 110 (Connection timed out))
23:09:33 Join midgey [0] (
23:10:11BHSPitMonkeyturn off hold and/or charge it
23:10:14SliMMpetit-lama-tout: is it charged
23:10:40dionoealinuxstb, jhMikeS : is it ok if i commit the subtitles code right away? or would you like to see a diff first?
23:11:38petit-lama-tout-now it start
23:11:41petit-lama-tout-very strange
23:12:45 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
23:12:50linuxstbdionoea: I'm curious to see a diff...
23:12:51jhMikeSdionoea: does it get into the frame rendering? I've got a change on that coming.
23:12:52TrueJournalsdionoea: subtitles code... ?
23:12:55***Saving seen data "./dancer.seen"
23:13:05 Join miepchen^schlaf [0] (
23:13:11dionoeajhMikeS: it adds a function call in the frame rendering thing
23:13:14jhMikeSthe frame dropping is pretty smooth overall now
23:13:18dionoeai'll post a diff online
23:15:54 Join Thundercloud_ [0] (
23:16:28dionoeaThis code supposes that 1/ the subs file is valid and 2/ it all fits into the buffer (the second supposition is stupid...)
23:17:01dionoeaBut it's a poc (kind of)
23:17:02 Join rift [0] (
23:17:06peturthe first one too imho
23:17:24dionoeawell validating everything would take way too much time
23:17:33peturthat's true yes
23:17:46peturas long as it doesn't crash...
23:17:53dionoeaI never said that :)
23:18:28jhMikeSED hovers from 14-16 fps on X5 now regulated...still a bit rough due to lack of B frames in the clip
23:18:36 Quit hannesd_ (Read error: 110 (Connection timed out))
23:18:40linuxstbdionoea: Why not simply allocate the buffer based on the filesize of the .srt file?
23:19:05dionoeabtw, you can get the subs here:
23:19:22dionoealinuxstb: i could do that too ... would be easier
23:19:41dionoeasubs file never take more that 200 kB right ? (and that'd be for a full movie)
23:20:08linuxstbI've just created an srt file from a 1-hour video and it's 39KB...
23:20:28dionoeaWe'll i'll do that then
23:20:49dionoeaIf you can have a look at the general aspects of the patch (like where i call the different functions and stuff) it'd be great
23:21:28peturBagder: something is wrong with the updating of the frontpage. The build of 20:56 did something strange it seems
23:21:46BagderI poked on it manually to trigger it
23:21:56Bagdersomething strange happened before that
23:22:21Bagderthe front page should show the dates and times correct of the actual commits
23:22:23pixelmanow you can see all commits when you look at the diffs
23:22:30dionoearb->filesize( ) returns the full file size ?
23:22:54linuxstbdionoea: I'm just commenting on things as I see them.... You call subs_render() even if there are no subtitles?
23:23:05Topydoes anyone know if anyone ever worked on or even suggested a realtime spectrum analysis on the recording screen? i am just listening to a few recordings i made and i think that would REALLY help setting up the mics in a noisy envoirement (where listening into the signal with headphones is impossible)
23:23:21dionoeayeah. I could move the check out of the function
23:23:24Topyvery low resolution, lets say 5 bars or so, would be enough
23:23:27 Quit Rob2222 ()
23:24:39 Quit funky ("leaving")
23:24:46Topybecause even with well isolating headphones (as my modded koss plugs) the outside influence is just to high to properly set up the mics, its always been a lot of guesswork aiming them towards the tops...
23:25:08 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
23:25:14 Quit petur ("here today, gone tomorrow")
23:25:28 Join miepchen^schlaf [0] (
23:25:54linuxstbdionoea: Why "#define atoi(a) rb->atoi(a)" ?
23:26:03dionoeafor my eyes
23:26:16pixelmabagawk: hmm... that came out strange- when you want to look at the last build (to see the details) from the builds page you will see all the commits from rev1 up to today (huge and interesting list btw)
23:26:19 Quit Thundercloud (Read error: 110 (Connection timed out))
23:26:23dionoealinuxstb: I can remove it if it's a problem
23:26:34pixelmathat was to Bagder... sorry
23:26:35 Quit davina ("byeeeeee!")
23:26:49jhMikeSI really think a general overlay solution should be in place
23:27:13jhMikeSand 44100 is also assumed I notice...
23:27:13 Part TrueJournals
23:27:26Bagderpixelma: oh, hehe
23:27:26dionoeaYep. Is that defined somewhere ?
23:27:33Bagdermy silly script...
23:27:59jhMikeSyes...but the video decoder decodes the HH:MM:SS on the frames too :)
23:28:42linuxstbdionoea: I think it's a general principle not to #define when it's not needed... Also I think "typedef struct" is frowned upon as well (but others will correct me)
23:28:43dionoeajhMikeS: well we could have a full overlay system. With a bunch of sub pictures to render at every time. But that would need more memory (i'm not sure about CPU time)
23:29:02dionoealinuxstb: ok. I'll change that.
23:29:17dionoeawould it be ok if I inline instead ?
23:29:49jhMikeSWe don't need anything but a mask
23:30:37linuxstbdionoea: inline what?
23:31:00dionoeainline int atoi( const char *s ){ return rb->atoi( s ); }
23:31:17linuxstbWhy not just use rb->atoi() directly?
23:31:18jhMikeSI'd really like to have YUV stuff not in the core since I plan on multiple blit functions to always use an optimized one.
23:31:23dionoeajhMikeS: you'd still need to render to the mask and then blend the mask on the video
23:32:05dionoeawhile I'm rendering directly to the video buffer
23:32:15jhMikeSThe mask only tells me which to use: lcd_framebuffer or yuv buffer ... in another sense, a _selectable_ chroma key may work too
23:32:23dionoeaOn targets with extra CPU it's a good idea to do it your way. I'm not sure for others
23:32:33 Join webguest35jaime [0] (i=522368cd@gateway/web/cgi-irc/
23:33:32webguest35jaimewondered if anyone new if rockbox will likely become available for samsung models
23:33:39jhMikeSYeah, for the really slow stuff, yeah, just cut the video away. I'd like this stuff handled automatically in the video out though
23:33:48 Quit petit-lama-tout- (Read error: 110 (Connection timed out))
23:34:00 Join MrStaticVoid [0] (
23:34:11linuxstbdionoea: Is your code rendering the text repeatedly each frame?
23:34:26dionoeayes. (since each video frame changes)
23:34:39dionoeaif you're rendering on the black bands it only updates on changes
23:34:57 Quit Mmmm (Remote closed the connection)
23:35:00linuxstbYes, but the text doesn't... You could render the text into a buffer (cache), and then just blit that buffer into the video frame.
23:35:02dionoea(but i don't remove the old text in that scenario ...)
23:35:51dionoeahum ... ok :) jhMikeS and linuxstb win :D I'll have an overlay cache which i'll blend onto the video
23:35:51 Quit webguest35jaime (Client Quit)
23:36:02dionoea(well, on the video y plane at least)
23:36:28jhMikeSlet us remind ourselves...what is "slow" not may not be for very long
23:37:40amiconnAnyone wanting to check out my newest win32 simulator unicode patch?
23:38:17amiconnNow it also prints unicode text correctly for debug messages, provided the console is set to use a truetype font
23:38:43 Quit billytwowilly (Read error: 110 (Connection timed out))
23:42:53 Quit Juice^ ("Leaving")
23:45:29SliMMamiconn: i'll do it, but tommorow
23:46:35 Join SliMM_ [0] (n=chatzill@
23:47:17 Quit SliMM_ (Client Quit)
23:47:56 Join drumr [0] (
23:48:10drumrim having some trouble with my ipod 5g
23:48:35drumri installed a new build on top of an old one, and when i try to play music (i tested mp3 and flac) it gives a codec error.
23:48:45drumrim going to completely replace the build now
23:48:50drumrbut, will that fix it?
23:48:59drumrand also, how do i clear the ram?
23:49:19drumri remember on the h3xx you hit the hold switch after it started up but before rockbox booted
23:49:27drumrnow that boots into the stock apple f/w
23:50:10 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
23:50:11linuxstbDo you have a rockbox.ipod file in both the root of your ipod and inside the .rockbox directory?
23:50:22 Quit amiconn (Nick collision from services.)
23:50:22 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
23:51:38 Join x-spec-t [0] (n=nwheeler@ubuntu/member/spec)
23:51:42 Quit Spec (Read error: 54 (Connection reset by peer))
23:51:45 Join ctaf [0] (
23:51:51XavierGrwhat's the fuss about removing the battery estimation feature?
23:51:57 Part ctaf
23:52:12XavierGrfor some works like a charm and in most cases it doesn't work right for non calibrated targets
23:52:51XavierGranyway I don't think that by removing it rockbox will gain much in binary size, so for the people that are annoyed by it, just don't use it :P
23:53:13*linuxstb doesn't see any point in removing it
23:53:31XavierGrit is a hate it or love it feature
23:56:57 Join Rob2222 [0] (
23:58:22XavierGron the other hand I can understand Llorean's frustration at always having to explain that this one won't work reliably for many targets, but IMHO I don't see a reason to remove it cause it will confuse some new users.
23:58:25 Join Spec[x] [0] (

Previous day | Next day